Artificial Intelligence-Based Control Method and System for Machining Centers

Abstract

This invention discloses a machining center control method and system based on artificial intelligence, belonging to the field of intelligent control technology. The method includes: creating a simulation model to generate a first trajectory of the milling head; optimizing the machining sequence and corresponding switching trajectory based on the path location and shortest distance principle of the milling head switching between different machining positions; real-time acquisition of sensing information, wear information, workpiece material, and cutting speed of the milling head, inputting cutting parameters into the cutting model, and generating a predicted cutting trajectory based on the simulation model; selecting a representative position in the predicted cutting trajectory and determining whether correction is needed; if correction is needed, calculating the correction position based on the principle of maximizing the projected area, and generating a second trajectory based on the correction position, the predicted cutting trajectory, and the optimized switching trajectory; and driving machining based on the control parameters of the second trajectory. This system effectively improves machining accuracy and efficiency through dynamic prediction and local trajectory correction.

Description

Technical Field

[0001] This invention relates to the field of intelligent control technology, and in particular to a control method and system for machining centers based on artificial intelligence. Background Technology

[0002] With the continuous development of the manufacturing industry, the requirements for the precision and efficiency of machining centers are becoming increasingly higher. Traditional machining center control methods mainly rely on preset programs and fixed trajectory planning, which are difficult to adapt to various complex situations that occur during the machining process and have certain limitations. Similar prior art includes Chinese Patent Publication No. CN113885433A, which proposes a digital control method and device for an intelligent milling machine, including: monitoring the control data of the intelligent milling machine in response to an activation signal, the control data including the machining control time of any machining device at a target position by the control unit of the intelligent milling machine; generating machining path data based on the control data, the machining path data being the first machining information of all machining devices participating in the machining in the intelligent milling machine from the moment the activation signal is received to the current moment, the first machining information including position information and time information; saving the machining path data at preset time intervals; when the intelligent milling machine reaches a first preset condition, retrieving the most recently saved machining path data, comparing the machining path data with the preset path data to obtain second machining information; and controlling the machining device of the intelligent milling machine to work based on the second machining information. This method can achieve intelligent control and realize the mid-processing of semi-finished products. Furthermore, a similar prior art exists in Chinese Patent Publication No. CN119439883A, which proposes an intelligent CNC machining method and system for mold manufacturing. This method includes acquiring the mold pattern to be cut and planning the milling cutter feed path, dividing it into multiple milling cutter feed sub-paths, determining the milling cutter control trajectory composed of spline curves, clustering the milling cutter control trajectories to obtain multiple trajectory sets, and milling the mold according to the milling cutter priority order. During milling, real-time machining trajectories are collected, curvature deviation values ​​are calculated, and if they exceed a preset value, the milling cutter control trajectory is corrected. Milling cutter priorities are sorted by working radius, and suitable trajectory sets are selected sequentially for milling until all trajectory sets are completed. Through intelligent path planning and real-time deviation correction, the efficiency and machining accuracy of milling cutter usage are significantly improved, ensuring high quality and high stability in mold manufacturing. While both of these patents address machine tool machining control issues, they fail to effectively optimize the switching trajectory of the milling head between different machining positions, resulting in long idle paths and low machining efficiency. Moreover, due to the lack of an effective real-time monitoring and correction mechanism during the processing, it is difficult to detect and correct any processing deviations in a timely manner, thus affecting the final processing quality. Summary of the Invention

[0003] This invention provides a machining center control method and system based on artificial intelligence, which significantly improves the control accuracy and efficiency of gantry machining centers through artificial intelligence technology. The method includes:

[0004] A simulation model is created based on the equipment information of the gantry machining equipment, the machining information of each machining position of the workpiece, the milling head information, and the simulation model to simulate the first trajectory of the milling head corresponding to each machining position.

[0005] The processing sequence and the corresponding switching trajectory in the first trajectory are optimized based on the switching path position and shortest distance principle of the milling head between different processing positions;

[0006] The real-time sensor information and wear information of the milling head, the material of the workpiece and the cutting shape are collected and input into the cutting model to obtain cutting parameters. The cutting parameters are then input into the simulation model to generate the predicted cutting trajectory in the first trajectory of the milling head.

[0007] Select representative positions in the predicted cutting trajectory, and determine whether the representative positions need to be corrected based on the polygon formed by each representative position and its adjacent representative positions;

[0008] When correction is required, the correction position of the representative position is obtained according to the principle of maximizing the graphic projection area, and a second trajectory is generated based on the correction position, the predicted cutting trajectory, and the switching trajectory.

[0009] The workpiece is processed based on the control parameters corresponding to the second trajectory, and the method of repeatedly acquiring the second trajectory is repeated when the real-time error is greater than the error threshold.

[0010] As a preferred embodiment of the present invention, the acquisition of the first trajectory of the milling head includes:

[0011] A simulation model is established based on the equipment information of the gantry machining equipment, the shape and material of the workpiece to be machined, and the positional relationship between the two. The machining information, milling head information, and equipment information corresponding to each machining position on the workpiece are input into the simulation model. The working process of the milling head is simulated according to the preset machining program corresponding to the machining information of each machining position to obtain the first trajectory of the milling head.

[0012] As a preferred embodiment of the present invention, optimizing the processing sorting and the switching trajectory corresponding to the first trajectory includes:

[0013] Based on the switching trajectory of the milling head between the first and second processing positions, and the simulation model of the state information of the path positions after processing, the first and second shortest distances corresponding to the path positions before and after processing are obtained based on the shortest path principle. When the first shortest distance is greater than the second shortest distance, the processing order of the path positions is prioritized over the processing order of the first and second processing positions; otherwise, it remains unchanged.

[0014] The switching trajectory in the first trajectory is optimized based on the trajectory corresponding to the minimum value between the first shortest distance and the second shortest distance.

[0015] As a preferred embodiment of the present invention, the acquisition of the predicted cutting trajectory includes:

[0016] During the milling process where the milling head processes the workpiece based on the control parameters corresponding to the first trajectory, the sensing information and wear information of the milling head are collected in real time. The sensing information, wear information, material of the workpiece, and cutting shape are input into the cutting model to obtain cutting information, which includes the rotation speed, number of rotations, feed speed, and cutting amount of the milling head. The cutting information is then input into the simulation model to obtain the predicted cutting trajectory corresponding to a future preset time period.

[0017] As a preferred embodiment of the present invention, determining whether the representative position needs correction includes:

[0018] The predicted cutting trajectory is divided into multiple small regions, and the center point of each small region is used as the machining position of the milling head. Any position of the milling head is used as the first point, and a point in the predicted cutting trajectory whose angle with the corresponding tangent plane of the i-th point is greater than a set angle is selected as the (i+1)-th point. The i-th point is used as the representative position, and the i-th point is connected to other adjacent points, forming a non-intersecting polygon. The polygon is mapped onto the tangent plane of the i-th point to obtain the mapped shape. The ratio of the first area of ​​the mapped shape to the second area of ​​the polygon is calculated. When the ratio is greater than or equal to a set value, the i-th point does not need to be corrected; when the ratio is less than the set value, the i-th point needs to be corrected.

[0019] As a preferred embodiment of the present invention, the generation of the second trajectory includes:

[0020] When the representative position needs to be corrected, the representative position is adjusted to meet the processing conditions while maximizing the mapping area of ​​the graphic corresponding to the adjusted representative position on the cross-section of the face where the representative position is located. The adjusted representative position is then used as the correction position, replacing the representative position in the predicted cutting trajectory. The replaced predicted cutting trajectory and the switching trajectory are then used as the second trajectory.

[0021] As a preferred technical solution of the present invention, the cutting model is a model trained with historical data, which includes historical sensing information of the milling head, historical wear information, workpiece material, cutting shape and corresponding historical cutting information.

[0022] As a preferred embodiment of the present invention, the gantry machining equipment includes multiple milling heads, and different milling heads can be automatically switched.

[0023] The present invention also provides an artificial intelligence-based machining center control system for implementing the above-described method, the system comprising:

[0024] The modeling unit is used to create a simulation model based on the equipment information of the gantry machining equipment, the machining information of each machining position of the workpiece, the milling head information, and the simulation model to simulate the first trajectory of the milling head corresponding to each machining position.

[0025] The optimization unit is used to optimize the processing sequence and the corresponding switching trajectory in the first trajectory based on the switching path position and the shortest distance principle between different processing positions of the milling head;

[0026] The prediction unit is used to input the real-time sensor information and wear information of the milling head, the material and cutting shape of the workpiece to the cutting model to obtain cutting parameters, and input the cutting parameters into the simulation model to generate the predicted cutting trajectory in the first trajectory of the milling head;

[0027] The judgment unit is used to select a representative position in the predicted cutting trajectory and determine whether the representative position needs to be corrected based on the polygon formed by each representative position and its adjacent representative positions.

[0028] The generation unit is used to obtain the correction position of the representative position according to the principle of maximizing the graphic projection area when correction is required, and generate a second trajectory based on the correction position, the predicted cutting trajectory and the switching trajectory;

[0029] The control unit is used to process the workpiece based on the control parameters corresponding to the second trajectory, and to repeatedly acquire the second trajectory when the real-time error is greater than the error threshold.

[0030] The present invention also provides a computer-readable storage medium storing instructions that, when executed by a processor, implement the method described above.

[0031] The beneficial effects of this invention are as follows:

[0032] This invention significantly improves the control accuracy and efficiency of gantry machining centers through artificial intelligence technology. A simulation model built from equipment, machining, and milling head information generates the aforementioned first trajectory corresponding to each machining position. Cutting parameters are dynamically updated using real-time sensor data and wear information, ensuring the predicted cutting trajectory always matches the actual state of the physical system, thus solving the problem of the disconnect between theoretical models and real-time working conditions in traditional CNC machining. Secondly, the unique "projection area ratio" correction mechanism represents position correction. By analyzing the ratio of the projected area of ​​the polygon formed by discrete points on the predicted trajectory on the cutting plane, it intelligently identifies the points requiring correction and optimizes spatial pose with the goal of maximizing the projected area, reducing machining errors on complex curved surfaces, while also reducing radial cutting forces and extending tool life. Furthermore, the trajectory optimization algorithm dynamically adjusts the machining sequence and shortens the idle path by simulating and predicting spatial changes before and after machining the path position. Finally, the system adopts a "local correction instead of global replanning" strategy to meet the real-time requirements of high-speed machining, forming a closed-loop control of "sensor monitoring - AI prediction - dynamic correction," which significantly improves the machining quality stability of precision parts such as large structural components. Attached Figure Description

[0033] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0034] Figure 1 This is a flowchart of the artificial intelligence-based machining center control method in an embodiment of the present invention;

[0035] Figure 2 This is a flowchart illustrating the method for optimizing processing sorting and switching trajectories in an embodiment of the present invention;

[0036] Figure 3 This is a flowchart of a method for determining whether a representative position needs correction in an embodiment of the present invention;

[0037] Figure 4 This is a structural diagram of the artificial intelligence-based machining center control system in an embodiment of the present invention. Detailed Implementation

[0038] This invention provides a machining center control method and system based on artificial intelligence. The terms "first," "second," "third," "fourth," etc. (if present) in the specification, claims, and accompanying drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" or "having" and any variations thereof are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0039] For ease of understanding, the specific steps of the embodiments of the present invention are described below, such as... Figure 1 As shown in the figure, an artificial intelligence-based machining center control method in this embodiment of the invention includes:

[0040] Step S1: Create a simulation model based on the equipment information of the gantry machining equipment, the machining information of each machining position of the workpiece, the milling head information, and the simulation model to simulate the first trajectory of the milling head corresponding to each machining position;

[0041] Specifically, the equipment information of the aforementioned gantry machining equipment includes the machine tool model, shape, and structure of the gantry machining equipment. Based on the aforementioned equipment information of the gantry machining equipment, the shape and material of the workpiece to be machined, and the positional relationship between the gantry machining equipment and the workpiece to be machined, a simulation model is established. The aforementioned machining information corresponding to each machining position on the workpiece to be machined, the milling head information corresponding to each machining position, and the aforementioned equipment information are input into the simulation model. The aforementioned machining information includes the machining shape and cutting amount of the machining position, and the aforementioned milling head information includes the size and material of the milling head. Combined with a preset machining program, an initial machining trajectory, namely the aforementioned first trajectory, is generated. The first trajectory includes the cutting path and the switching path between different machining positions. The aforementioned technical solution can obtain a high-precision simulation model and obtain the first trajectory of the milling head corresponding to each machining position based on the aforementioned simulation model, laying the foundation for subsequent high-precision and high-efficiency machining of the workpiece to be machined.

[0042] Step S2: Optimize the processing sequence and the corresponding switching trajectory in the first trajectory based on the switching path position and shortest distance principle of the milling head between different processing positions;

[0043] Specifically, for multi-position machining scenarios, an optimization mechanism based on "path position state prediction" is proposed. For example, when the milling head switches from position A to position B, if the path area C is the area to be machined, the spatial morphology of area C after material removal is simulated through a simulation model, such as the height change after milling the boss, and the shortest collision-free path before and after material removal is calculated: the current path length L1 is compared with the path length L2 after machining area C. If L1>L2, the machining sequence of area C is advanced, and the physical characteristic of "material removal creating space" is used to shorten the idle time; at the same time, the switching trajectory is updated with the trajectory corresponding to the minimum value of L1 and L2. This optimization shortens the idle time path, which significantly improves efficiency, especially for complex box-type parts, and avoids the defect of the traditional TSP (Traveling Salesman Problem) algorithm ignoring the dynamic environmental changes during machining.

[0044] Step S3: Input the real-time sensor information, wear information, material of the workpiece and cutting shape of the milling head into the cutting model to obtain cutting parameters, and input the cutting parameters into the simulation model to generate the predicted cutting trajectory in the first trajectory of the milling head;

[0045] Specifically, during the initial trajectory execution, multi-source real-time data is collected synchronously, such as the temperature, friction, vibration data, and wear information of the milling head. This data is input into a cutting model trained on historical machining data, such as an LSTM neural network, which outputs dynamic cutting information, including the rotational speed, rotational speed, feed rate, and cutting amount of the milling head. The model compensates for nonlinear factors such as stiffness attenuation and localized material hardening caused by tool wear through online learning. The updated parameters drive the simulation model to regenerate the predicted cutting trajectory for future time periods, such as the next 200ms, achieving a transition from a "preset theoretical path" to a "condition-adaptive path," thereby reducing machining accuracy fluctuations by more than 30%.

[0046] Step S4: Select a representative position in the predicted cutting trajectory, and determine whether the representative position needs to be corrected based on the polygon formed by each representative position and its adjacent representative positions;

[0047] Specifically, the predicted trajectory is discretized into a dense point cloud, that is, the continuous predicted cutting trajectory is divided into multiple small regions. The center point of each region is regarded as the theoretical machining point of the milling head, and a representative position, i.e., the i-th point, is selected. It is ensured that the selected point is located in the region where the curvature of the trajectory changes, but the change is not too large. The i-th point is connected with its adjacent points to form a non-intersecting polygon. The polygon is projected onto the tangent plane corresponding to the i-th point to generate a two-dimensional mapping figure. The ratio of the area of ​​the mapping figure, i.e., the first area S1, to the area of ​​the original polygon, i.e., the second area S2, is calculated. When the ratio is greater than or equal to the set value, no correction is required. Conversely, when the ratio is less than the set value, it indicates that there is a high degree of distortion or steep undulation in the local surface. The tool will cause overcutting or undercutting if it follows the original path, and correction needs to be triggered. The above technical solution breaks through the limitations of traditional curvature threshold-based methods. By quantifying the "developability of the surface" through area ratio, the accuracy of overcutting / undercutting identification in freeform surface machining is improved to 92%.

[0048] Step S5: When correction is required, obtain the correction position of the representative position according to the principle of maximizing the projected area of ​​the graphic, and generate a second trajectory based on the correction position, the predicted cutting trajectory and the switching trajectory;

[0049] Specifically, for the representative position that needs to be corrected, an optimization problem is solved within a constrained space. The constrained space can be the tool-workpiece collision envelope or the machine tool travel limit. The spatial coordinates of the representative position are adjusted to maximize the mapping area of ​​the corresponding polygon on the corresponding tangent surface of the representative position, and the correction position is obtained. The correction point replaces the aforementioned representative position in the original trajectory, while the trajectory in the non-corrected area remains unchanged. Finally, it is combined with the optimized switching trajectory to form a second trajectory. This local correction strategy keeps the calculation time within 5ms and avoids real-time interruptions caused by global replanning. The above technical solution reduces the machining error of complex curved surfaces while ensuring machining efficiency.

[0050] Step S6: Process the workpiece based on the control parameters corresponding to the second trajectory, and repeat the method of obtaining the second trajectory when the real-time error is greater than the error threshold.

[0051] Specifically, the second trajectory is post-processed into machine tool executable instructions, such as speed look-ahead control and acceleration smoothing, to drive multi-milling head collaborative machining. The system continuously monitors the deviation between the actual trajectory and the second trajectory. When the error exceeds the error threshold, the re-prediction-S5 correction loop, which starts from step S3, is triggered, i.e., the above-mentioned method of repeatedly obtaining the second trajectory. This closed-loop architecture integrates "feedforward prediction" (step S3) and "feedback correction" (steps S4-S5) to form a real-time control loop with a small response delay, thereby reducing the range of accuracy fluctuations throughout the long-cycle machining of large structural parts.

[0052] Furthermore, the acquisition of the first trajectory of the milling head includes:

[0053] A simulation model is established based on the equipment information of the gantry machining equipment, the shape and material of the workpiece to be machined, and the positional relationship between the two. The machining information, milling head information, and equipment information corresponding to each machining position on the workpiece are input into the simulation model. The working process of the milling head is simulated according to the preset machining program corresponding to the machining information of each machining position to obtain the first trajectory of the milling head.

[0054] Specifically, the equipment information of the aforementioned gantry machining equipment includes the machine tool model, shape, and structure of the gantry machining equipment. Based on the aforementioned equipment information of the gantry machining equipment, the shape and material of the workpiece to be machined, and the positional relationship between the gantry machining equipment and the workpiece to be machined, a simulation model is established. The aforementioned machining information corresponding to each machining position on the workpiece to be machined, the milling head information corresponding to each machining position, and the aforementioned equipment information are input into the simulation model. The aforementioned machining information includes the machining shape and cutting amount of the machining position to be machined, and the aforementioned milling head information includes the size and material of the milling head. The working process of each milling head is simulated by each simulation model based on the preset machining program corresponding to the machining information of each position, and the first trajectory of each milling head is obtained. The aforementioned preset machining program is the control program corresponding to the milling head. The aforementioned first trajectory includes the switching trajectory between different machining positions and the cutting trajectory at the machining position. The aforementioned technical solution can obtain the ideal machining path of the milling head corresponding to each machining position, i.e., the aforementioned first path.

[0055] Furthermore, the processing order and the corresponding switching trajectory in the first trajectory are optimized, such as... Figure 2 As shown, it includes:

[0056] Based on the switching trajectory of the milling head between the first and second processing positions, and the simulation model of the state information of the path positions after processing, the first and second shortest distances corresponding to the path positions before and after processing are obtained based on the shortest path principle. When the first shortest distance is greater than the second shortest distance, the processing order of the path positions is prioritized over the processing order of the first and second processing positions; otherwise, it remains unchanged.

[0057] The switching trajectory in the first trajectory is optimized based on the trajectory corresponding to the minimum value between the first shortest distance and the second shortest distance.

[0058] Specifically, when the milling head switches from the first processing position to the second processing position for processing, the path of the switching trajectory between the first and second processing positions may be the processing position. Before processing, the switching trajectory may need to detour around the path position. Therefore, processing at the path position may shorten the switching trajectory, for example, due to spatial changes caused by material removal. Therefore, the simulation model simulates the state after processing at the path position, i.e., the state after removing excess material at the path position, and calculates the first and second shortest distances based on the shortest distance principle. The shortest distance principle is that when the milling head switches from the first processing position... When switching to the second processing position, a preset distance needs to be maintained between the milling head and the surface of the workpiece. This minimizes the path the milling head takes when switching from the first processing position to the second processing position. If the first shortest distance is greater than the second shortest distance, it indicates that processing at the aforementioned path position can shorten the path. Therefore, the processing order of the path position is set before the first and second processing positions. Conversely, the processing order of the path position remains after the first and second processing positions. The switching trajectory is optimized based on the trajectory corresponding to the minimum of the first and second shortest distances. This technical solution can obtain a shorter switching trajectory, thereby improving processing efficiency.

[0059] Furthermore, the acquisition of the predicted cutting trajectory includes:

[0060] During the milling process where the milling head processes the workpiece based on the control parameters corresponding to the first trajectory, the sensing information and wear information of the milling head are collected in real time. The sensing information, wear information, material of the workpiece, and cutting shape are input into the cutting model to obtain cutting information, which includes the rotation speed, number of rotations, feed speed, and cutting amount of the milling head. The cutting information is then input into the simulation model to obtain the predicted cutting trajectory corresponding to a future preset time period.

[0061] Specifically, since the milling head's performance changes with working conditions during operation, affecting cutting accuracy, the milling head, in the process of machining the workpiece based on the control parameters corresponding to the first trajectory, captures the mechanical properties of the milling head by real-time acquisition of its sensor information and wear information. This sensor information includes the milling head's temperature, friction, vibration data, and wear information. The sensor information, wear information, and the material and cutting shape of the workpiece are then input into the cutting model. The cutting information, including the milling head's rotational speed, number of rotations, feed rate, and cutting amount (the number of rotations corresponds to a single feed), is input into the simulation model to obtain the predicted cutting trajectory for a future preset time period. This technical solution dynamically acquires the predicted cutting trajectory based on the milling head's real-time state, and in conjunction with the above steps, improves both cutting efficiency and cutting accuracy.

[0062] Further, determine whether the representative position needs correction, such as... Figure 3 As shown, it includes:

[0063] The predicted cutting trajectory is divided into multiple small regions, and the center point of each small region is used as the machining position of the milling head. Any position of the milling head is used as the first point, and a point in the predicted cutting trajectory whose angle with the corresponding tangent plane of the i-th point is greater than a set angle is selected as the (i+1)-th point. The i-th point is used as the representative position, and the i-th point is connected to other adjacent points, forming a non-intersecting polygon. The polygon is mapped onto the tangent plane of the i-th point to obtain the mapped shape. The ratio of the first area of ​​the mapped shape to the second area of ​​the polygon is calculated. When the ratio is greater than or equal to a set value, the i-th point does not need to be corrected; when the ratio is less than the set value, the i-th point needs to be corrected.

[0064] Specifically, the continuous predicted cutting trajectory is divided into multiple small regions, with the center point of each region considered as the theoretical machining point of the milling head. This discretization transforms the complex three-dimensional path into a set of points that can be quantified and analyzed. An arbitrary starting point is taken as the i-th point, representing the position. Based on spatial geometric constraints, adjacent points with an angle greater than a set threshold to the tangent plane (normal plane) of that point are selected as the (i+1)-th point. This set threshold can be 10°, ensuring that the selected points are located in areas where the curvature of the trajectory changes, but not excessively. These areas may be due to machining deviations caused by vibrations of the milling head or gantry milling equipment. Connecting the i-th point with its adjacent points forms a non-intersecting polygon (usually a triangle or quadrilateral). This polygon is a near representation of the local curved surface of the workpiece. The geometric representation projects a polygonal mesh onto the tangent plane (ideal machining plane) corresponding to the i-th point, generating a two-dimensional mapping graphic. The ratio of the area of ​​the mapping graphic (the first area S1) to the area of ​​the original polygonal graphic (the second area S2) is calculated and compared with a set value. When the ratio is greater than or equal to the set value, it indicates that the actual surface closely matches the ideal tangent plane and no correction is needed. Conversely, when the ratio is less than the set value, it indicates that the local surface has high distortion or steep undulations, and the milling head will overcut or undercut if it follows the original path, requiring correction. This technical solution can accurately identify whether there is uneven cutting at each of the above representative positions, facilitating subsequent calibration and laying the foundation for improving cutting accuracy.

[0065] Furthermore, the generation of the second trajectory includes:

[0066] When the representative position needs to be corrected, the representative position is adjusted to meet the processing conditions while maximizing the mapping area of ​​the graphic corresponding to the adjusted representative position on the cross-section of the face where the representative position is located. The adjusted representative position is then used as the correction position, replacing the representative position in the predicted cutting trajectory. The replaced predicted cutting trajectory and the switching trajectory are then used as the second trajectory.

[0067] Specifically, when the projected area ratio (mapped area / original polygon area) of the representative position (i.e., the i-th point) is lower than a set value, it indicates that the cutting at that point may have produced errors, resulting in insufficient smoothness. This triggers a correction mechanism. The correction unit, centered on the representative position, iteratively adjusts the three-dimensional coordinates of that point within a spatial range that satisfies machining constraints (such as tool length, workpiece boundary, collision avoidance, etc.). The optimization objective is to maximize the mapped area of ​​the polygon formed by the representative position and its adjacent points on the current cutting plane. The optimized corrected position replaces the representative position in the original predicted trajectory and is integrated with the switching trajectory generated by the optimization unit to generate a second trajectory that can directly drive the equipment. This process retains the trajectory data of the non-corrected sections and only reconstructs the local path. This technical solution significantly improves machining quality while ensuring real-time performance. Its technical essence is to optimize the spatial pose to make the milling head cutting state approach the ideal effect.

[0068] Furthermore, the cutting model is a model trained with historical data, which includes historical sensing information of the milling head, historical wear information, workpiece material, cutting shape, and corresponding historical cutting information.

[0069] Furthermore, the gantry machining equipment includes multiple milling heads, and different milling heads can be automatically switched.

[0070] This invention also provides an artificial intelligence-based machining center control system for implementing the above-described method, such as... Figure 4 As shown, the system includes:

[0071] The modeling unit is used to create a simulation model, which simulates the first trajectory of each milling head based on the equipment information of the gantry machining equipment, the machining information of each machining position of the workpiece, the milling head information, and the simulation model.

[0072] The optimization unit is used to optimize the processing sequence and the corresponding switching trajectory in the first trajectory based on the switching path position and the shortest distance principle between different processing positions of the milling head;

[0073] The prediction unit is used to input the real-time sensor information and wear information of the milling head, the material and cutting shape of the workpiece to the cutting model to obtain cutting parameters, and input the cutting parameters into the simulation model to generate the predicted cutting trajectory in the first trajectory of the milling head;

[0074] The judgment unit is used to select a representative position in the predicted cutting trajectory and determine whether the representative position needs to be corrected based on the polygon formed by each representative position and its adjacent representative positions.

[0075] The correction unit is used to obtain the correction position of the representative position according to the principle of maximizing the graphic projection area when correction is required, and to generate a second trajectory based on the correction position, the predicted cutting trajectory and the switching trajectory.

[0076] The control unit is used to process the workpiece based on the control parameters corresponding to the second trajectory, and to repeatedly acquire the second trajectory when the real-time error is greater than the error threshold.

[0077] The present invention also provides a computer-readable storage medium storing instructions that, when executed by a processor, implement the method described above.

[0078] In summary, this invention significantly improves the control accuracy and efficiency of gantry machining centers through artificial intelligence technology. The simulation model, established using equipment information, machining information, and milling head information, generates the aforementioned first trajectory corresponding to each machining position. Cutting parameters are dynamically updated using real-time sensor data and wear information, ensuring the predicted cutting trajectory always matches the actual state of the physical system, thus solving the problem of the disconnect between theoretical models and real-time working conditions in traditional CNC machining. Secondly, the unique "projection area ratio" correction mechanism represents position correction. By analyzing the ratio of the projected area of ​​the polygon formed by discrete points on the predicted trajectory on the cutting plane, it intelligently identifies the points requiring correction and optimizes spatial pose with the goal of maximizing the projected area, reducing machining errors on complex curved surfaces, while also reducing radial cutting forces and extending tool life. Furthermore, the trajectory optimization algorithm dynamically adjusts the machining sequence and shortens the idle path by simulating and predicting spatial changes before and after machining the path position. Finally, the system adopts a "local correction instead of global replanning" strategy to meet the real-time requirements of high-speed machining, forming a closed-loop control of "sensor monitoring - AI prediction - dynamic correction," which significantly improves the machining quality stability of precision parts such as large structural components.

[0079] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0080] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0081] The above-described embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. An artificial intelligence-based machining center control method, characterized by, The method comprises: Creating a simulation model, simulating a first trajectory of each to-be-processed position corresponding to the milling head based on equipment information of the gantry machining equipment, machining information of each to-be-processed position of the to-be-processed piece, milling head information and the simulation model; Optimizing machining order and switching trajectory in the first trajectory according to switching approach position and shortest distance principle of the milling head between different to-be-processed positions; Inputting real-time collected sensing information and wear information of the milling head, material and cutting shape of the to-be-processed piece into a cutting model to obtain cutting parameters, inputting the cutting parameters into the simulation model to generate a predicted cutting trajectory in the first trajectory of the milling head; Selecting representative positions in the predicted cutting trajectory, and judging whether the representative positions need to be corrected according to a polygon formed between each representative position and adjacent representative positions; When correction is needed, obtaining a correction position of the representative position according to the maximum projection area principle, generating a second trajectory based on the correction position, the predicted cutting trajectory and the switching trajectory; Based on the control parameters corresponding to the second trajectory, machining the to-be-processed piece, and when real-time error is greater than an error threshold, repeating the method of obtaining the second trajectory.

2. The method of claim 1, wherein, The generation of the motion trajectory of each milling head and the related control instructions comprises: Based on the equipment information of the gantry machining equipment, the shape and material of the to-be-processed piece and the positional relationship therebetween, a simulation model is established, the machining information corresponding to each to-be-processed position, the milling head information and the equipment information are input into the simulation model, and the working process of each milling head is simulated according to the preset program corresponding to the machining information of each to-be-processed position, to obtain a first path of each milling head.

3. The method of claim 1, wherein, Optimizing machining order and switching trajectory in the first trajectory comprises: According to the switching trajectory of the milling head between the first to-be-processed position and the second to-be-processed position, and according to the state information of the approach position of the simulated switching trajectory after machining, and based on the shortest path principle, the first shortest distance and the second shortest distance corresponding to the approach position before and after machining are obtained, when the first shortest distance is greater than the second shortest distance, the machining order of the approach position is prior to the machining order of the first to-be-processed position and the second to-be-processed position, otherwise, it remains unchanged; According to the trajectory corresponding to the minimum value of the first shortest distance and the second shortest distance, the switching trajectory in the first trajectory is optimized.

4. The method of claim 1, wherein, The acquisition of the predicted cutting trajectory comprises: During the machining process of the to-be-processed piece by the milling head based on the control parameters corresponding to the first trajectory, the sensing information and wear information of the milling head are collected in real time, and the sensing information, wear information, material and cutting shape of the to-be-processed piece are input into the cutting model to obtain cutting information, the cutting information includes the rotating speed, rotating number, feeding speed and cutting amount of the milling head, and the cutting information is input into the simulation model to obtain the predicted cutting trajectory corresponding to a future preset time period.

5. The method of claim 1, wherein, The judgment of whether the representative positions need to be corrected comprises: The predicted cutting trajectory is divided into a plurality of small regions, and the center point of the small region is taken as the machining position of the milling head, any position of the milling head is taken as the first site, and the point in the predicted cutting trajectory corresponding to the selected surface of the i-th site is taken as the i+1-th site, the i-th site is taken as the representative position, and the i-th site and the adjacent other sites are connected to form a non-intersecting polygonal figure, the polygonal figure is mapped onto the surface of the i-th site, the mapping figure is obtained, the ratio of the first area of the mapping figure to the second area of the polygonal figure is calculated, and when the ratio is greater than or equal to a set value, the i-th site does not need to be corrected, and when the ratio is less than the set value, the i-th site needs to be corrected.

6. The method of claim 1, wherein, The generation of the second trajectory includes: When the representative position needs to be corrected, the representative position is adjusted while meeting the machining conditions, so that the adjusted representative position corresponds to the maximum mapping area of the figure on the surface of the representative position, the adjusted representative position is taken as the corrected position, the corrected position replaces the representative position in the predicted cutting trajectory, and the predicted cutting trajectory after replacement and the switching trajectory are taken as the second trajectory.

7. The method of claim 1, wherein, The cutting model is a model trained by historical data, and the historical data includes historical sensing information of the milling head, historical wear information, workpiece material, cutting shape, and corresponding historical cutting information.

8. The method of claim 1, wherein, The gantry machining equipment includes a plurality of milling heads, and the milling heads can be automatically switched between different milling heads.

9. Artificial intelligence-based machining center control system for implementing the method according to any one of claims 1-8, characterized in that, The system includes: A modeling unit is configured to create a simulation model, simulate a first trajectory of each machining position corresponding to the milling head based on equipment information of the gantry machining equipment, machining information of each machining position of a workpiece to be machined, milling head information, and the simulation model; An optimization unit is configured to optimize machining sequencing and a switching trajectory in the first trajectory according to switching approach positions of the milling head between different machining positions and a shortest distance principle; A prediction unit is configured to input real-time sensing information, wear information of the milling head, material and cutting shape of the workpiece to be machined into a cutting model to obtain cutting parameters, and input the cutting parameters into the simulation model to generate a predicted cutting trajectory in the first trajectory of the milling head; A judgment unit is configured to select a representative position in the predicted cutting trajectory, and determine whether the representative position needs to be corrected according to a polygonal figure formed between each representative position and adjacent representative positions; A generation unit is configured to, when correction is needed, obtain a corrected position of the representative position according to a maximum projection area principle of the figure, and generate a second trajectory based on the corrected position, the predicted cutting trajectory, and the switching trajectory; A control unit is configured to machine the workpiece to be machined based on control parameters corresponding to the second trajectory, and repeat the method of obtaining the second trajectory when a real-time error is greater than an error threshold.

10. A computer-readable storage medium having stored thereon instructions, the instructions comprising, The instructions are executed by the processor to implement the method of any one of claims 1-8. The instructions are executed by the processor to implement the method of any one of claims 1-8.

Images