Numerical control intelligent programming system and method based on AI large model

Abstract

The invention provides an AI large model-based numerical control intelligent programming system and method. The method comprises the following steps of: determining a stability characteristic value of each machining stage in a numerical control machining process through a cutting force characteristic and a vibration spectrum characteristic of numerical control machining; when the stability characteristic value exceeds a safety threshold value in the numerical control machining process, performing characteristic comparison on a workpiece shape model machined by the numerical control machine tool and an original design model of a target workpiece to obtain deformation deviation distribution of each machining area in the workpiece shape model; determining a deviation program segment of each control module in the initial numerical control program through all deformation deviation distribution, and further determining deviation characteristics of each deviation program segment; and based on each deviation characteristic, driving the AI large model to perform deviation correction on each deviation program segment in the initial numerical control program to obtain an optimized numerical control program of the target workpiece. Based on the scheme, cooperative feedback correction of multiple physical quantities in the machining process can be achieved, and therefore the dynamic stability of a numerical control program can be improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of numerical control programming, more particularly, the present application relates to an AI large model-based numerical control intelligent programming system and method. BACKGROUND

[0002] Numerical control programming is the process of converting a three-dimensional model of a part into a set of instructions that can be recognized by a numerical control machine. Programmers use numerical control programming software to plan tool paths, select cutting parameters, and generate program code that controls the movement of the machine. This code drives the tool to perform cutting, drilling, and milling operations. Numerical control programming is a key step in achieving digital manufacturing and automated processing.

[0003] The programming and verification process of traditional computer-aided manufacturing systems relies heavily on geometric simulation. This simulation mode uses Boolean operations to simulate material removal and checks whether the tool path will cause overcutting, undercutting, or static interference between the tool, tool holder, and workpiece or fixture. However, purely geometric verification cannot simulate physical effects during processing, such as assessing whether the cutting force exceeds the load limit of the machine or tool, predicting process system chatter caused by improper parameters, or determining the negative impact of vibration on surface quality during cutting. As a result, a seemingly perfect program on the geometric level may fail due to dynamic problems in actual physical processing. Therefore, how to achieve coordinated feedback correction of multiple physical quantities during processing to improve the dynamic stability of numerical control programs has become a challenge in the industry. SUMMARY

[0004] The present application provides an AI large model-based numerical control intelligent programming system and method, which can achieve coordinated feedback correction of multiple physical quantities during processing, thereby improving the dynamic stability of numerical control programs.

[0005] In a first aspect, the present application provides an AI large model-based numerical control intelligent programming method, comprising: When using an initial numerical control program to program and simulate a target workpiece, the numerical control machining process of the target workpiece is simulated and verified in the virtual machining environment of the programming simulation. At the same time, physical state information during dynamic simulation and verification is collected, and a workpiece shape model after machining by a numerical control machine is obtained. Based on the geometric level relationship of non-machining components in the numerical control machine under the linkage of all motion axes, cutting force features and vibration spectrum features of numerical control machining are extracted from the physical state information, and then stability feature values of each machining stage in the numerical control machining process are determined through the cutting force features and the vibration spectrum features. comparing the workpiece shape model machined by the numerical control machine with the original design model of the target workpiece to obtain a deformation deviation distribution of each machining area in the workpiece shape model, determining a deviation program segment of each control module in the initial numerical control program through all the deformation deviation distributions, and further determining a deviation feature of each deviation program segment; driving the AI large model based on each deviation feature to correct the deviation of each deviation program segment in the initial numerical control program to obtain a numerical control program of the optimized target workpiece.

[0006] In some embodiments, the cutting force feature and the vibration spectrum feature of the numerical control machining are extracted from the physical state information based on the geometric hierarchical relationship of the non-machining components in the numerical control machine under the linkage of all motion axes, and specifically include: a kinematic chain three-dimensional model including a machine bed, a spindle, a tool turret, and each linear axis and rotary axis is established based on the geometric hierarchical relationship of the non-machining components in the numerical control machine under the linkage of all motion axes; synchronizing and aligning the physical state information with the machining state in the kinematic chain three-dimensional model to obtain cutting force signals and vibration acceleration signals of each machining stage in the numerical control machining; extracting the cutting force feature of the numerical control machining from all the cutting force signals; performing frequency domain conversion on each vibration acceleration signal to obtain a frequency spectrum of each machining stage in the numerical control machining, and further extracting the vibration spectrum feature of the numerical control machining from all the frequency spectra.

[0007] In some embodiments, the stability feature value of each machining stage in the numerical control machining process is determined by the cutting force feature and the vibration spectrum feature, and specifically includes: for each machining stage in the numerical control machining process, extracting a cutting force-based stability index of the machining stage from the cutting force feature; extracting a vibration-based stability index of the machining stage from the vibration spectrum feature; combining the cutting force-based stability index and the vibration-based stability index into a stability feature value of the machining stage, and further determining the stability feature value of each machining stage in the numerical control machining process.

[0008] In some embodiments, the workpiece shape model machined by the numerical control machine is compared with the original design model of the target workpiece to obtain a deformation deviation distribution of each machining area in the workpiece shape model, and specifically includes: dividing the workpiece shape model machined by the numerical control machine into multiple machining areas; aligning the workpiece shape model machined by the numerical control machine with the original design model of the target workpiece in three-dimensional space to obtain a three-dimensional deviation field of each machining area. generate a deformation deviation distribution of each processing area in the workpiece shape model through all three-dimensional deviation fields.

[0009] In some embodiments, determining the deviation program segment of each control module in the initial numerical control program through all deformation deviation distributions specifically includes: obtaining a mapping relationship between the workpiece processing area and the program control module; filtering out the program segment corresponding to each deformation deviation distribution based on the mapping relationship; filtering out the deviation program segment of each control module in the initial numerical control program from all program segments.

[0010] In some embodiments, determining the deviation feature of each deviation program segment specifically includes: for each deviation program segment, obtaining the tool path and process parameters in the deviation program segment; comparing the tool path with the standard tool path of the control module to obtain the path deviation of the control module corresponding to the deviation program segment; comparing the process parameters with the standard process parameters of the control module to obtain the process feature deviation of the control module corresponding to the deviation program segment; determining the deviation feature of the deviation program segment through the path deviation and the process feature deviation, and thereby obtaining the deviation feature of each deviation program segment.

[0011] In some embodiments, driving the AI large model to correct the deviation of each deviation program segment in the initial numerical control program based on each deviation feature to obtain the numerical control program of the optimized target workpiece specifically includes: initializing an AI large model based on deep learning; taking each deviation feature as an input feature vector in the AI large model; taking each deviation program segment in the initial numerical control program as an entity to be optimized in the AI large model; using the AI large model to iteratively correct and optimize the numerical control program of the target workpiece to obtain the numerical control program of the optimized target workpiece.

[0012] In a second aspect, the present application provides an AI large model-based numerical control intelligent programming system, which comprises a programming optimization unit, and the programming optimization unit comprises: an acquisition module, configured to, when programming simulation is performed on a target workpiece using an initial numerical control program, simulate and verify the numerical control machining process of the target workpiece in a virtual machining environment of the programming simulation, collect physical state information in the dynamic simulation verification, and obtain a workpiece shape model after machining of a numerical control machine tool; The processing module is configured to extract a cutting force feature and a vibration spectrum feature of numerical control machining from the physical state information based on a geometric hierarchical relationship of a non-machining component in the numerical control machine tool under linkage of all motion axes, and further determine a stability feature value of each machining stage in the numerical control machining process through the cutting force feature and the vibration spectrum feature. The processing module is further configured to perform feature comparison between a workpiece shape model machined by the numerical control machine tool and an original design model of a target workpiece when the stability feature value exceeds a safety threshold of the numerical control machining process, to obtain a deformation deviation distribution of each machining area in the workpiece shape model, to determine a deviation program segment of each control module in the initial numerical control program through all of the deformation deviation distributions, and to further determine a deviation feature of each deviation program segment. The execution module is configured to drive the AI large model to perform deviation correction on each deviation program segment in the initial numerical control program based on the deviation feature, to obtain a numerical control program of an optimized target workpiece.

[0013] In a third aspect, the present application provides a computer device, which comprises a memory and a processor, the memory is configured to store a computer program, and the processor is configured to call and run the computer program from the memory, so that the computer device executes the AI large model based numerical control intelligent programming method described above.

[0014] In a fourth aspect, the present application provides a computer readable storage medium, which stores instructions or codes, when the instructions or codes are run on a computer, the computer is caused to execute the AI large model based numerical control intelligent programming method described above.

[0015] The technical scheme provided by the embodiments of the present application has the following beneficial effects: The application provides an AI large model-based numerical control intelligent programming system and method. When an initial numerical control program is used to program and simulate a target workpiece, the numerical control machining process of the target workpiece is simulated and verified in a virtual machining environment of the programming simulation, physical state information in dynamic simulation and verification is collected, and a workpiece shape model after machining of a numerical control machine tool is obtained. Based on a geometric hierarchical relationship of non-machining components in the numerical control machine tool under linkage of all motion axes, cutting force features and vibration spectrum features of numerical control machining are extracted from the physical state information, and then stability feature values of each machining stage in the numerical control machining process are determined through the cutting force features and the vibration spectrum features. When the stability feature values exceed a safety threshold value of the numerical control machining process, the workpiece shape model after machining of the numerical control machine tool is compared with an original design model of the target workpiece in features, and deformation deviation distributions of each machining area in the workpiece shape model are obtained. Through all the deformation deviation distributions, deviation program segments of each control module in the initial numerical control program are determined, and then deviation features of each deviation program segment are determined. Based on the deviation features, an AI large model is driven to correct the deviation program segments in the initial numerical control program, and a numerical control program of an optimized target workpiece is obtained.

[0016] It can be seen that, in the present application, based on the deviation characteristics of each AI large model, the deviation of each deviation program segment in the initial numerical control program is corrected to obtain the numerical control program of the optimized target workpiece; first, the stability characteristic value is determined to obtain the quantitative criterion of the dynamic stability of the machining process, thereby providing the numerical control program with early warning and intervention ability ahead of the occurrence of geometric error, by extracting the cutting force characteristics and vibration spectrum characteristics from the physical state information and fusing them into a comprehensive stability characteristic value, the vibration and stress of the machine tool-tool-workpiece process system in the machining process are monitored in real time, and when the stability characteristic value exceeds the safety threshold, it indicates that the process system is at the critical point of instability, and chatter may occur or has occurred, at this time, although irreversible geometric deviation has not been caused, the machining quality has been seriously threatened, the stability characteristic value as a leading indicator provides a key intervention window period for the numerical control program, so that it can trigger the subsequent diagnosis and correction process before the geometric deviation actually occurs, realizing the fundamental change from passive detection to active prevention, and thereby improving the dynamic stability of the numerical control program; then, the deviation characteristics are determined to obtain the accurate diagnosis report of the cause and position of the machining defect, thereby realizing the accurate tracing and targeted correction from macroscopic deviation phenomenon to microscopic program instruction, the specific machining area is located through the deformation deviation distribution in the deviation characteristics, and then the deviation program segment causing the deviation of the area is traced to the deviation program segment, thereby extracting the specific deviation characteristics of the deviation program segment on the tool path and process parameters, converting the abstract unstable or out-of-tolerance phenomenon into a specific and operable engineering technical problem description, and then taking the deviation characteristics as the input of the AI large model, the AI large model can generate highly targeted optimization strategies (for example: only optimizing the feed rate of a specific corner), thereby realizing the targeted correction of the initial numerical control program, while correcting the defects, the reasonable part of the original program is maximized, and the dynamic stability and machining precision of the final output program are efficiently and reliably improved; in summary, based on the above scheme, the coordinated feedback correction of multiple physical quantities in the machining process can be realized, thereby improving the dynamic stability of the numerical control program. BRIEF DESCRIPTION OF DRAWINGS

[0017] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiments or prior art description. Obviously, the drawings in the following description are only some embodiments of the present application, and those skilled in the art can also obtain other drawings according to these drawings without creative labor.

[0018] Figure 1 is an exemplary flowchart of the AI large model-based numerical control intelligent programming method according to some embodiments of the present application; Figure 2is a flowchart of a determination bias procedure segment according to some embodiments of the present application; Figure 3 is a structural diagram of a programming optimization unit according to some embodiments of the present application; Figure 4 is a structural diagram of a computer device for implementing an AI large model-based numerical control intelligent programming method according to some embodiments of the present application. DETAILED DESCRIPTION

[0019] In order to better understand the technical solutions of the present application, the technical solutions of the present application will be described in detail below in conjunction with the drawings in the specification and specific embodiments.

[0020] Reference Figure 1 The figure is an exemplary flowchart of an AI large model-based numerical control intelligent programming method according to some embodiments of the present application, which mainly includes the following steps: In step 101, when using an initial numerical control program to perform programming simulation on a target workpiece, the numerical control machining process of the target workpiece is simulated and verified in a virtual machining environment of the programming simulation, and physical state information in the dynamic simulation verification is collected, and a workpiece shape model after machining of the numerical control machine tool is obtained.

[0021] It should be noted that in the present application, the initial numerical control program is a set of machining instructions for controlling the motion trajectory and machining action of the numerical control machine tool, and the initial numerical control program is a numerical control code automatically generated by computer-aided manufacturing software according to a three-dimensional model of a part without practical verification; the virtual machining environment is a software platform integrating a machine tool kinematics model, a geometric model, a physical engine and a rendering engine; the physical state information is state data representing stability and load conditions in the machining process during dynamic simulation, which includes cutting force (reflecting the load borne by the tool), vibration spectrum (reflecting the stability of the machining system, used for early warning of chatter), spindle torque (reflecting the cutting resistance) and tool-workpiece contact state; the workpiece shape model is a three-dimensional geometric model of the workpiece as a test of whether the machining result is qualified.

[0022] In a specific implementation, first, a built-in machine tool model library, a tool library and a fixture library are called to configure a kinematic model of a virtual numerical control machine tool according to physical parameters (for example, a stroke range, a spindle interface and an axle system structure) of an actual machining center, so as to import and position a three-dimensional design model of a target workpiece to a clamping position of a virtual workbench, and generate an initial blank model according to a blank size, thereby completing the construction of a digital mapping environment consistent with a real production scene, and taking the constructed digital mapping environment as a virtual machining environment for programming simulation; then, an initial numerical control program is loaded into the virtual machining environment, a code instruction is parsed line by line by an interpretation module of a virtual machining environment kernel, a material removal process is simulated after driving each motion axis of the virtual machine tool to be linked, a physical calculation module integrated in the material removal process dynamically calculates cutting force, vibration spectrum, spindle torque and a tool-workpiece contact state at each moment according to a tool path, cutting parameters and material properties, and the cutting force, the vibration spectrum, the spindle torque and the tool-workpiece contact state are taken as physical state information in dynamic simulation verification; finally, after the dynamic simulation is completed, a machined workpiece shape model is calculated by a material removal algorithm of the virtual machining environment.

[0023] In step 102, cutting force features and vibration spectrum features of numerical control machining are extracted from the physical state information based on a geometric hierarchical relationship of non-machining components in the numerical control machine tool under linkage of all motion axes, and then stability feature values of each machining stage in the numerical control machining process are determined through the cutting force features and the vibration spectrum features.

[0024] In some embodiments, the extraction of the cutting force features and the vibration spectrum features of the numerical control machining from the physical state information based on the geometric hierarchical relationship of the non-machining components in the numerical control machine tool under linkage of all motion axes can be implemented by the following steps: A three-dimensional model of a motion chain including a machine bed, a spindle, a tool turret, each linear axis and a rotary axis is established based on the geometric hierarchical relationship of the non-machining components in the numerical control machine tool under linkage of all motion axes; The physical state information is synchronized and aligned with a machining state in the three-dimensional model of the motion chain, to obtain cutting force signals and vibration acceleration signals of each machining stage in the numerical control machining; Cutting force features of the numerical control machining are extracted from all the cutting force signals; Each vibration acceleration signal is subjected to frequency domain conversion to obtain a frequency spectrum of each machining stage in the numerical control machining, and then vibration spectrum features of the numerical control machining are extracted from all the frequency spectra.

[0025] It should be noted that in the present application, the cutting force feature is a feature for characterizing the core state of the cutting process; the vibration spectrum feature is a feature for locating the vibration source and identifying whether harmful chatter has occurred in the machining process; the kinematic chain three-dimensional model is a three-dimensional digital assembly used to reflect the connection relationship and motion transmission logic between the mechanical components of the numerical control machine tool, which accurately defines the geometric constraints and linkage relationship between all moving parts, ensures that the virtual environment can calculate the real-time accurate position of all non-machining parts of the machine tool at any position and attitude of the cutter, and provides a spatial context environment for subsequent physical simulation and collision detection; the cutting force signal is a signal reflecting the change in load size when the cutter contacts the workpiece; the vibration acceleration signal is a signal reflecting the degree of vibration intensity of the machine tool structure due to force.

[0026] In specific implementation, first, based on the mechanical assembly drawing of the numerical control machine tool and the kinematics principle, the three-dimensional model of each linear axis and rotary axis is added and adjusted in the simulation software starting from the fixed bed body, and assembled strictly according to the actual parent-child hierarchical relationship, and finally a complete kinematic chain from the bed body to the spindle, then to the tool turret and the cutter is formed, so that the complete kinematic chain is assembled to the three-dimensional model as the kinematic chain three-dimensional model; secondly, while driving the cutter to cut in the kinematic chain three-dimensional model, the built-in physical engine calculates the simulated cutting force and vibration response according to the material properties of the cutter and the workpiece, the cutting parameters and the real-time material cutting volume, the simulation core is to stamp the simulated physical signal and the machine tool motion state with a unified time stamp, ensuring that the cutting force signal and the vibration acceleration signal at each moment strictly correspond to the machining state such as the cutting position and the cutting depth of the cutter at that time, so that the cutting force signal and the vibration acceleration signal in each machining stage of numerical control machining can be obtained; then, for each machining stage, the cutting force signal of the machining stage is filtered to remove noise interference, the average value of all cutting forces in the cutting force signal is calculated as the average load value, the root mean square value of all cutting forces in the cutting force signal is calculated as the sustained level value, and the peak value of all cutting forces in the cutting force signal is calculated as the impact load value, and the set of average load value, sustained level value and impact load value is taken as the value range of the stage feature value of the cutting force in the machining stage, so that the stage feature value of the cutting force in the machining stage can be obtained, and the set of all stage feature values is taken as the cutting force feature of numerical control machining through the above-mentioned manner; finally, the fast Fourier transform algorithm is used to convert it from the time dimension to the frequency dimension, and a frequency spectrum graph with frequency as the horizontal axis and vibration intensity as the vertical axis is generated, the sharp peaks with amplitude significantly higher than the background noise are identified from the frequency spectrum graph, and the frequency value and the corresponding amplitude size of each sharp peak are recorded, so that the set of all frequency values and the corresponding amplitude sizes is taken as the vibration spectrum feature of numerical control machining.

[0027] In some embodiments, the determination of the stability characteristic value of each machining stage in the numerical control machining process by the cutting force feature and the vibration frequency spectrum feature can be achieved by the following steps: For each machining stage in the numerical control machining process, extracting a cutting force-based stability index of the machining stage from the cutting force feature; Extracting a vibration-based stability index of the machining stage from the vibration frequency spectrum feature; Fusing the cutting force-based stability index and the vibration-based stability index into a stability characteristic value of the machining stage, and then determining the stability characteristic value of each machining stage in the numerical control machining process.

[0028] It should be noted that in this application, the stability characteristic value is a quantitative value for determining the stability degree of the machining stage; the cutting force-based stability index is a numerical index for evaluating the smoothness of the machining process by quantifying the fluctuation of the cutting force signal; and the vibration-based stability index is a numerical index for evaluating the stability of the dynamic behavior of the machining by quantifying the intensity of the vibration signal.

[0029] In specific implementation, firstly, for each machining stage in the numerical control machining process, the average load value, the sustained level value and the impact load value of the machining stage are obtained from the cutting force feature, so as to calculate the weighted sum of the average load value, the sustained level value and the impact load value as the cutting force-based stability index by using a preset weight; then, the frequency value and the corresponding amplitude size of each peak are obtained from the vibration frequency spectrum feature, and the sum of the squares of the amplitude size corresponding to each frequency value is calculated as the total vibration energy in the machining stage; in order to avoid too large absolute value difference under different working conditions, the ratio of the total vibration energy to a safety reference value set according to the type of machine tool and tool can be used as a relative index representing the deviation degree of the current vibration level from the safety state, so as to use the relative index as the vibration-based stability index in the machining stage; finally, the sum of the cutting force-based stability index and the vibration-based stability index is calculated as the stability characteristic value of the machining stage; in other embodiments, in order to improve the calculation accuracy of the stability characteristic value, the fusion weight of the cutting force and the vibration can be personalized according to the different types of target workpieces, so as to calculate the weighted sum of the cutting force-based stability index and the vibration-based stability index as the stability characteristic value of the machining stage, which is not limited here. The stability characteristic value of each machining stage in the numerical control machining process can be obtained by the above-mentioned manner.

[0030] In step 103, when the stability characteristic value exceeds the safety threshold of the numerical control machining process, the workpiece shape model after machining of the numerical control machine is compared with the original design model of the target workpiece in features, to obtain a deformation deviation distribution of each machining area in the workpiece shape model, and through all the deformation deviation distributions, a deviation program segment of each control module in the initial numerical control program is determined, and then a deviation feature of each deviation program segment is determined.

[0031] It should be noted that in the present application, when the stability characteristic value exceeds the safety threshold of the numerical control machining process, it indicates that the comprehensive dynamic performance of the current machining stage has tended to be unstable, and the abnormal fluctuation of cutting force or the intensification of mechanical vibration is significantly reducing the machining quality and safety. At this time, the program deviation correction must be started immediately, so as to associate the over-standard stability characteristic value to the specific code segment in the numerical control program that causes chatter or load mutation through reverse mapping, and according to the dominant frequency of the vibration spectrum and the peak value distribution of the cutting force fed back by the physical simulation, the adaptive control system is driven to compensate and optimize the feed rate, spindle speed or tool path trajectory of the suspected code segment online, so as to inhibit the chatter source and smooth the cutting load through parameter re-planning under the condition of not stopping, so as to make the process system dynamic characteristics return to the stable domain again.

[0032] In some embodiments, comparing the workpiece shape model after machining of the numerical control machine with the original design model of the target workpiece in features to obtain the deformation deviation distribution of each machining area in the workpiece shape model can be realized by the following steps: Dividing the workpiece shape model after machining of the numerical control machine into multiple machining areas; Aligning the workpiece shape model after machining of the numerical control machine with the original design model of the target workpiece in three-dimensional space to obtain a three-dimensional deviation field of each machining area; Generating the deformation deviation distribution of each machining area in the workpiece shape model through all the three-dimensional deviation fields.

[0033] It should be noted that in the present application, the deformation deviation distribution is a quantitative value for finally judging whether the machining quality of each machining area in the workpiece shape model is qualified; the machining area is a sub-regional area for positioning the error source in the workpiece shape model after machining of the numerical control machine; and the three-dimensional deviation field is a data set representing the normal deviation size and direction of each point on the machined surface relative to the corresponding point on the original design model.

[0034] In a specific implementation, first, a computer-aided recognition algorithm is used to automatically divide the workpiece shape model processed by the numerical control machine tool into a plurality of logically independent regions as processing regions, and a plurality of processing regions can be obtained by the above method. In other embodiments, the division basis includes different processing features (for example, planar region, cavity region, hole system region, free-form surface region), or different processing procedures (for example, rough machining region, semi-finish machining region, finish machining region), which are not limited here. Then, the workpiece shape model processed by the numerical control machine tool is accurately aligned with the original design model in the three-dimensional space, to ensure that the reference and coordinate system of the two are completely consistent, and the normal distance from each sampling point on the surface of the workpiece shape model to the nearest point on the surface of the original design model is calculated. If the workpiece shape model surface point is outside the original design model, it is a positive deviation (overcut), and vice versa, it is a negative deviation (undercut). For each processing region, arrange all the normal distances in the processing region according to the corresponding spatial position to obtain the three-dimensional deviation field of the processing region, and the three-dimensional deviation field of each processing region can be obtained by the above method. Finally, based on the boundary of each processing region in the workpiece shape model, all normal distances in the processing region are obtained from the three-dimensional deviation field of the processing region, so as to calculate the maximum deviation, average deviation and standard deviation of all normal distances in the processing region as the deformation deviation distribution of the processing region, and the deformation deviation distribution of each processing region in the workpiece shape model can be obtained by the above method.

[0035] In some embodiments, the deviation program segment of each control module in the initial numerical control program is determined by all the deformation deviation distributions, and the reference Figure 2 The figure is a flowchart for determining the deviation program segment in some embodiments of the present application. The deviation program segment in the present embodiment can be realized by the following steps: In step 1031, the mapping relationship between the workpiece processing region and the program control module is obtained. In step 1032, the program segment corresponding to each deformation deviation distribution is selected based on the mapping relationship. In step 1033, the deviation program segment of each control module in the initial numerical control program is selected from all the program segments.

[0036] It should be noted that in the present application, the deviation program segment is the numerical control code instruction that needs to be corrected in the initial numerical control program; the mapping relationship is the corresponding association relationship between each processing region on the workpiece and the program control module in the numerical control program responsible for cutting; and the program segment is a collection of all numerical control codes associated with the processing region with deformation deviation.

[0037] In the implementation, firstly, the mapping relationship between the workpiece processing area and the program control module is obtained from the central control station of the simulation software kernel; then, the program segment in the program control module corresponding to each deformation deviation distribution and the processing area is screened out from the mapping relationship as the program segment corresponding to the corresponding deformation deviation distribution, so that the program segment corresponding to each deformation deviation distribution is obtained; finally, for each control module in the initial numerical control program, the program segment belonging to the control module is screened out from all the program segments as the deviation program segment of the control module, so that the deviation program segment of each control module in the initial numerical control program is obtained through the above manner.

[0038] In some embodiments, the determination of the deviation feature of each deviation program segment can be implemented by the following steps: For each deviation program segment, the tool path and the process parameter in the deviation program segment are obtained; The tool path is compared with the standard tool path of the control module to obtain the path deviation of the control module corresponding to the deviation program segment; The process parameter is compared with the standard process parameter of the control module to obtain the process feature deviation of the control module corresponding to the deviation program segment; The deviation feature of the deviation program segment is determined through the path deviation and the process feature deviation, and then the deviation feature of each deviation program segment is obtained.

[0039] It should be noted that in the present application, the deviation feature, the tool path, the process parameter, the path deviation, and the process feature deviation.

[0040] In a specific implementation, first, for each deviation program segment, the geometric motion instructions (for example, G01 linear interpolation, G02 / G03 circular arc interpolation) of the deviation program segment are extracted by parsing the numerical control code of the deviation program segment, and the continuous motion trajectory of the tool center point is calculated as the tool path by the post-processing kernel of the software; at the same time, the spindle speed, feed speed and cutting depth used when the deviation program segment is executed are directly read from the code of the deviation program segment or derived by calculation, and the set of spindle speed, feed speed and cutting depth is taken as the machining parameters of the deviation program segment; secondly, the tool path is compared with the standard tool path automatically generated by the computer-aided manufacturing system according to the ideal geometric model, the normal distance between the corresponding points on the two paths is calculated, and the maximum value, average value and root mean square value of all normal distances are calculated, and the weighted sum of the maximum value, average value and root mean square value is taken as the path deviation of the corresponding control module of the deviation program segment, wherein the weight of the weighted fusion can be preset through historical experience; then, the process parameters are compared with the standard parameter range recommended in the built-in process knowledge base for the current machining material, tool type and machining type, the relative deviation percentage of the actual value of each parameter in the process parameters to the standard value is calculated as the deviation value of the corresponding parameter, for example, the percentage of the actual feed speed exceeding the recommended value range, and the weighted sum of all deviation values is calculated as the process feature deviation, wherein the weight of the weighted fusion can be preset through historical experience; finally, the path deviation and process feature deviation are taken as inputs, and a comprehensive analysis is performed according to the preset rule base, for example, if the path deviation of the program segment shows that the trajectory at the corner is not smooth, and the process feature deviation value shows that the feed rate at the corner is too high, then the deviation feature is summarized as "corner impact under high feed rate", and a specific description is output for each deviation program segment as the deviation feature of the deviation program segment, and the deviation features of each deviation program segment can be obtained through the above-mentioned manner.

[0041] In step 104, the AI large model is driven based on each deviation feature to correct the deviation of each deviation program segment in the initial numerical control program, and the numerical control program of the optimized target workpiece is obtained.

[0042] In some embodiments, the AI large model can be driven based on each deviation feature to correct the deviation of each deviation program segment in the initial numerical control program, and the numerical control program of the optimized target workpiece can be realized by the following steps: Initialize an AI large model based on deep learning; Take each deviation feature as an input feature vector in the AI large model; Take each deviation program segment in the initial numerical control program as an entity to be optimized in the AI large model; Use the AI large model to iteratively correct and optimize the numerical control program of the target workpiece, and obtain the numerical control program of the optimized target workpiece.

[0043] It should be noted that in this application, the AI large model is a deep learning network trained by a large number of numerical control machining cases and process rules. Through a large amount of historical machining data and process knowledge training, a complex nonlinear mapping from "machining problem features" to "numerical control program optimization strategy" is established. The AI large model receives the structured deviation feature vector (for example: overcut amount, vibration frequency) as input, extracts high-dimensional abstract features through multi-layer nonlinear transformation, and then identifies the deviation mode and the root cause; Subsequently, the AI large model generates specific parameter correction amount or path adjustment scheme according to the learned optimization rule library (for example: adjusting the feed rate to suppress chatter, modifying the tool path to avoid overcutting) for the associated to-be-optimized program segment entity, and through multiple simulation-evaluation-correction iteration cycles, finally outputs an optimized numerical control program that has been significantly improved in terms of geometric accuracy, dynamic stability and machining efficiency.

[0044] In addition, another aspect of the present application, in some embodiments, the present application provides an AI large model-based numerical control intelligent programming system, which comprises a programming optimization unit, which is described with reference to Figure 3 The figure is a structural schematic diagram of the programming optimization unit according to some embodiments of the present application, which comprises an acquisition module 201, a processing module 202 and an execution module 203, which are described as follows: The acquisition module 201 is mainly used for simulating and verifying the numerical control machining process of the target workpiece in the virtual machining environment of the programming simulation when using the initial numerical control program to program and simulate the target workpiece, collecting the physical state information in the dynamic simulation verification, and obtaining the workpiece shape model after machining by the numerical control machine tool; The processing module 202 is used for extracting the cutting force features and vibration spectrum features of numerical control machining from the physical state information based on the geometric hierarchical relationship of non-machining components in the numerical control machine tool under the linkage of all motion axes, and then determining the stability characteristic value of each machining stage in the numerical control machining process through the cutting force features and the vibration spectrum features; It should be noted that the processing module 202 is also used for comparing the workpiece shape model after machining by the numerical control machine tool with the original design model of the target workpiece when the stability characteristic value exceeds the safety threshold of the numerical control machining process, obtaining the deformation deviation distribution of each machining area in the workpiece shape model, and determining the deviation program segment of each control module in the initial numerical control program through all the deformation deviation distributions, and then determining the deviation features of each deviation program segment; The execution module 203 is mainly used for driving the AI large model to correct the deviation of each deviation program segment in the initial NC program based on each deviation feature, so as to obtain the NC program of the optimized target workpiece.

[0045] The above describes the examples of the AI large model-based NC intelligent programming system and method provided by the embodiments of the present application in detail. It can be understood that the corresponding device contains the hardware structure and / or software module corresponding to the execution of each function in order to realize the above functions. Those skilled in the art should easily realize that the units and algorithm steps of each example described in combination with the embodiments disclosed in the present application can be realized in the form of hardware or a combination of hardware and computer software. Whether a certain function is executed in the form of hardware or computer software driven hardware depends on the specific application and design constraints of the technical solution. Professional technicians can use different methods to realize the described functions for each specific application, but such implementation should not be considered beyond the scope of the present application.

[0046] In some embodiments, the present application also provides a computer device, which comprises a memory and a processor, the memory is used to store a computer program, and the processor is used to call and run the computer program from the memory, so that the computer device executes the AI large model-based NC intelligent programming method described above.

[0047] In some embodiments, with reference to Figure 4 The dashed line in the figure indicates that the unit or the module is optional, and the figure is a structural schematic diagram of a computer device for implementing the AI large model-based NC intelligent programming method according to the embodiments of the present application. The AI large model-based NC intelligent programming method described in the above embodiments can be implemented by the computer device shown in the figure, which comprises at least one processor 301, a memory 302 and at least one communication unit 305, and the computer device can be a terminal device or a server or a chip. Figure 4

[0048] The processor 301 can be a general-purpose processor or a special-purpose processor. For example, the processor 301 can be a central processing unit (CPU), which can be used to control the computer device, execute the software program, and process the data of the software program. The computer device can also include a communication unit 305 to realize the input (reception) and output (transmission) of signals.

[0049] For example, the computer device can be a chip, the communication unit 305 can be an input and / or output circuit of the chip, or the communication unit 305 can be a communication interface of the chip, and the chip can be a component of a terminal device or a network device or other device.​

[0050] For another example, the computer device can be a terminal device or a server, the communication unit 305 can be a transceiver of the terminal device or the server, or the communication unit 305 can be a transceiver circuit of the terminal device or the server.

[0051] The computer device can include one or more memories 302, on which a program 304 is stored, the program 304 can be run by the processor 301 to generate instructions 303, so that the processor 301 executes the method described in the above method embodiments according to the instructions 303. Alternatively, the memory 302 can also store data (such as a target audit model). Alternatively, the processor 301 can also read the data stored in the memory 302, and the data can be stored in the same storage address as the program 304, or the data can be stored in a different storage address from the program 304.

[0052] The processor 301 and the memory 302 can be separately arranged, or can be integrated together, for example, integrated on a system on chip (SOC) of the terminal device.

[0053] It should be understood that each step of the above method embodiments can be completed by a logic circuit in the form of hardware in the processor 301 or instructions in the form of software, and the processor 301 can be a CPU, a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA), or other programmable logic devices, for example, discrete gates, transistor logic devices, or discrete hardware components.

[0054] Those skilled in the art should understand that the embodiments of the present application can be provided as a method, a system, or a computer program product. Therefore, the present application can take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present application can take the form of a computer program product implemented on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROMs, optical storage, etc.) containing computer-usable program code.

[0055] For example, in some embodiments, the present application also provides a computer-readable storage medium, the computer-readable storage medium stores instructions or codes, when the instructions or codes are run on a computer, the computer executes the AI-based large model numerical control intelligent programming method described above.

[0056] While the preferred embodiments of the application have been described, additional variations and modifications can be made to these embodiments by those skilled in the art once they have the benefit of the present disclosure without departing from the spirit and scope of the application. Accordingly, it is intended that such additions and modifications be included within the scope of the application. It is the following claims, including any amendments thereto, which define the scope of the application.

[0057] Obviously, numerous modifications and variations of the present application are possible in light of the above teachings. It is therefore to be understood that within the scope of the appended claims and their equivalents, the application can be practiced otherwise than as specifically described.

Claims

1. An AI-based large model numerical control intelligent programming method, characterized in that, The method comprises the following steps: When programming simulation is performed on a target workpiece using an initial numerical control program, a numerical control machining process of the target workpiece is simulated and verified in a virtual machining environment of the programming simulation, physical state information in dynamic simulation and verification is collected, and a workpiece shape model after machining of a numerical control machine tool is obtained; Cutting force features and vibration spectrum features of numerical control machining are extracted from the physical state information based on geometric hierarchical relationships of non-machining components in the numerical control machine tool under linkage of all movement axes, and then stability feature values of each machining stage in the numerical control machining process are determined through the cutting force features and the vibration spectrum features; When the stability feature values exceed a safety threshold of the numerical control machining process, a feature comparison is performed between the workpiece shape model after machining of the numerical control machine tool and an original design model of the target workpiece, deformation deviation distributions of each machining region in the workpiece shape model are obtained, deviation program segments of each control module in the initial numerical control program are determined through all the deformation deviation distributions, and then deviation features of each deviation program segment are determined; Based on the deviation features, an AI large model is driven to correct the deviation program segments in the initial numerical control program, and a numerical control program of an optimized target workpiece is obtained.

2. The method of claim 1, wherein, The cutting force features and the vibration spectrum features of numerical control machining are extracted from the physical state information based on geometric hierarchical relationships of non-machining components in the numerical control machine tool under linkage of all movement axes, and the method comprises the following steps: A movement chain three-dimensional model containing a bed body, a spindle, a tool turret, and each linear axis and rotary axis is established based on geometric hierarchical relationships of non-machining components in the numerical control machine tool under linkage of all movement axes; The physical state information is synchronously aligned with a machining state in the movement chain three-dimensional model, and cutting force signals and vibration acceleration signals of each machining stage in numerical control machining are obtained; Cutting force features of numerical control machining are extracted from all the cutting force signals; Each vibration acceleration signal is subjected to frequency domain conversion, frequency spectrum diagrams of each machining stage in numerical control machining are obtained, and then vibration spectrum features of numerical control machining are extracted from all the frequency spectrum diagrams.

3. The method of claim 1, wherein, The stability feature values of each machining stage in the numerical control machining process are determined through the cutting force features and the vibration spectrum features, and the method comprises the following steps: For each machining stage in the numerical control machining process, a cutting force-based stability index in the machining stage is extracted from the cutting force features; A vibration-based stability index in the machining stage is extracted from the vibration spectrum features; The cutting force-based stability index and the vibration-based stability index are fused into a stability feature value of the machining stage, and then the stability feature values of each machining stage in the numerical control machining process are determined.

4. The method of claim 1, wherein, The feature comparison between the workpiece shape model after machining of the numerical control machine tool and the original design model of the target workpiece to obtain the deformation deviation distributions of each machining region in the workpiece shape model comprises the following steps: The workpiece shape model after machining of the numerical control machine tool is divided into multiple machining regions; The workpiece shape model after machining of the numerical control machine tool is aligned with the original design model of the target workpiece in three-dimensional space, and three-dimensional deviation fields of each machining region are obtained; The deformation deviation distributions of each machining region in the workpiece shape model are generated through all the three-dimensional deviation fields.

5. The method of claim 1, wherein, The specific steps of determining the deviation program segments of each control module in the initial numerical control program through all the deformation deviation distributions include: obtaining the mapping relationship between the workpiece machining area and the program control module; screening the program segments corresponding to each deformation deviation distribution based on the mapping relationship; screening the deviation program segments of each control module in the initial numerical control program from all the program segments.

6. The method of claim 1, wherein, The specific steps of determining the deviation characteristics of each deviation program segment include: for each deviation program segment, obtaining the tool path and process parameters in the deviation program segment; comparing the tool path with the standard tool path of the control module to obtain the path deviation of the control module corresponding to the deviation program segment; comparing the process parameters with the standard process parameters of the control module to obtain the process characteristic deviation of the control module corresponding to the deviation program segment; determining the deviation characteristics of the deviation program segment through the path deviation and the process characteristic deviation, and then obtaining the deviation characteristics of each deviation program segment.

7. The method of claim 1, wherein, The specific steps of driving the AI large model to correct the deviation of each deviation program segment in the initial numerical control program based on the deviation characteristics include: initializing an AI large model based on deep learning; taking each deviation characteristic as an input feature vector in the AI large model; taking each deviation program segment in the initial numerical control program as an entity to be optimized in the AI large model; using the AI large model to iteratively correct and optimize the numerical control program of the target workpiece to obtain the optimized numerical control program of the target workpiece.

8. An AI large model-based numerical control intelligent programming system, the AI large model-based numerical control intelligent programming system comprising a programming optimization unit, characterized in that, The programming optimization unit includes: an acquisition module, configured to, when programming simulation is performed on the target workpiece using the initial numerical control program, perform simulation verification on the numerical control machining process of the target workpiece in a virtual machining environment of the programming simulation, collect physical state information in the dynamic simulation verification, and obtain a workpiece shape model after machining by the numerical control machine tool; a processing module, configured to extract cutting force characteristics and vibration spectrum characteristics of the numerical control machining from the physical state information based on the geometric hierarchical relationship of non-machining components in the numerical control machine tool under linkage of all motion axes, and then determine stability characteristic values of each machining stage in the numerical control machining process through the cutting force characteristics and the vibration spectrum characteristics; the processing module is further configured to, when the stability characteristic values exceed a safety threshold of the numerical control machining process, perform feature comparison between the workpiece shape model after machining by the numerical control machine tool and an original design model of the target workpiece to obtain deformation deviation distributions of each machining area in the workpiece shape model, determine the deviation program segments of each control module in the initial numerical control program through all the deformation deviation distributions, and then determine the deviation characteristics of each deviation program segment; an execution module, configured to drive an AI large model to correct the deviation of each deviation program segment in the initial numerical control program based on the deviation characteristics, and obtain the optimized numerical control program of the target workpiece.

9. A computer device, comprising: The computer device includes a memory and a processor, the memory is used to store a computer program, and the processor is used to call and run the computer program from the memory, so that the computer device executes the AI large model-based numerical control intelligent programming method in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer readable storage medium stores instructions or codes, and when the instructions or codes run on the computer, the computer executes the AI large model-based numerical control intelligent programming method as claimed in any one of claims 1 to 7.

Images