Intelligent tool setting control method and system

Abstract

The invention relates to the technical field of mechanical detection, in particular to an intelligent tool setting control method and system, and the method comprises the steps: collecting burr morphology data and tool state parameters of a slitting edge; establishing a mapping relationship between burr generation and processing parameters based on a process DOE model, wherein the process DOE model is used for extracting a key factor set influencing the slitting quality; when the real-time monitoring value of the key factor set deviates from the optimization interval of the process DOE model, generating a position and posture compensation instruction of the slitting tool rest; iteratively updating the key factor weight distribution of the process DOE model; and when the variance of the key factor weight distribution is converged to a preset threshold value, determining that the tool setting control closed-loop optimization is completed. According to the method, the defects of dynamic response hysteresis and insufficient compensation precision in traditional tool setting control are effectively overcome, and closed-loop self-optimization control of the position and posture of the slitting tool rest-process parameters-burr suppression is constructed through a three-dimensional dynamic compensation mechanism of real-time fusion of multi-source sensing data.

Description

Technical Field

[0001] The present invention relates to the field of mechanical detection technology, and in particular to an intelligent tool setting control method and system. Background Art

[0002] In the field of precision slitting, especially when slitting ultra-thin materials such as lithium battery electrodes and polymer films, dynamic interactive control of the slitting tool holder and the material is a key factor in determining slitting quality and tool life. Traditional tool setting control methods are mostly based on preset fixed process parameter thresholds, setting key parameters such as slitting speed and tool pressure through manual experience or offline test data. However, in actual continuous processing, the coupling of multiple factors such as tool wear, material property fluctuations, and equipment vibration leads to continuous dynamic changes in the processing state. Traditional static control solutions make it difficult to achieve real-time coordinated optimization of slitting quality and tool status.

[0003] The main defect of the existing technology is the lack of a dynamic adaptability mechanism. On the one hand, traditional methods usually analyze the impact of a single process parameter on the slitting quality in isolation, and are unable to establish a dynamic mapping relationship between process parameters, tool health status and burr morphology. When the tool edge shows progressive wear or sudden chipping, it is difficult for the system to perceive and generate an effective compensation strategy in time, resulting in the burr suppression effect being significantly attenuated with the passage of processing time; on the other hand, the current compensation control mostly adopts a fixed rule trigger mechanism, which does not take into account dynamic factors such as the remaining life of the tool and the difference in material batches, and can easily lead to over-compensation or compensation lag problems, which in turn accelerates abnormal tool wear. Summary of the Invention

[0004] The present invention provides an intelligent tool setting control method and system, thereby effectively solving the problems pointed out in the background technology.

[0005] In order to achieve the above object, the technical solution adopted by the present invention is: An intelligent tool setting control method, comprising: Drive the slitting tool holder to cut the electrode, and simultaneously collect the burr shape data and tool status parameters of the slitting edge; Establishing a mapping relationship between burr generation and processing parameters based on a process DOE model, wherein the process DOE model is used to extract a set of key factors affecting slitting quality; When the real-time monitoring value of the key factor set deviates from the optimization range of the process DOE model, a posture compensation instruction of the slitting tool holder is generated; Adjusting the spatial posture of the slitting tool holder according to the posture compensation instruction, and iteratively updating the key factor weight distribution of the process DOE model based on the adjusted burr morphology data; When the variance of the key factor weight distribution converges to a preset threshold, the tool setting control closed-loop optimization is determined to be completed.

[0006] Furthermore, a mapping relationship between burr generation and processing parameters is established based on the process DOE model, including: The tool state parameters collected are used as coordinate axes to construct a three-dimensional process space, whose parameter range covers the entire tool life cycle and the fluctuation range of material properties; Mapping the real-time burr morphology data to a three-dimensional process space to generate a burr feature contour distribution map; The tool wear gradient, cutting edge chipping position and vibration energy level characteristics are marked in the three-dimensional process space to form a multi-dimensional correlation map of process parameters-burr morphology-tool status.

[0007] Furthermore, tool wear gradient, edge chipping location, and vibration energy level characteristics are marked in the three-dimensional process space to form a multi-dimensional correlation map of process parameters, burr morphology, and tool status, including: A nonlinear mapping relationship is established based on the historical data of the tool wear gradient and the slitting speed. A spatial cluster analysis is performed on the cutting edge chipping position parameters to generate a chipping sensitive area distribution map. The high-frequency vibration risk domain is marked in combination with the frequency domain energy partition result of the vibration energy level characteristics. The burr feature contour distribution map is stored in bins according to the material plastic deformation coefficient, and is dynamically associated with the tool health status data; A response surface model including slitting speed, tool pressure, material coefficient, three-dimensional process vector, maximum burr height and root stress concentration coefficient was constructed.

[0008] Furthermore, when the real-time monitoring value of the key factor set deviates from the optimization range of the process DOE model, a posture compensation instruction for the slitting tool holder is generated, including: Calculating the deviation of the key control factor in the three-dimensional process space based on the real-time value of the key control factor; generating an initial compensation parameter set by combining the burr feature contour distribution map in the multi-dimensional correlation map with the tool health status data; The initial compensation parameter set is screened for feasibility, and the posture compensation instruction that satisfies both the burr suppression goal and the tool life protection is output.

[0009] Furthermore, generating the initial compensation parameter set includes: According to the burr feature contour distribution map, the direction with the maximum gradient change rate is selected as the priority compensation axis; Calling the wear gradient in the tool health status data to calculate the compensation weight coefficient; The initial compensation parameter set is generated on the priority compensation axis according to the compensation weight coefficient.

[0010] Furthermore, the calculation formula of the compensation weight coefficient is: ; in, is the compensation weight coefficient, is the tool wear gradient, The maximum allowable tool wear threshold.

[0011] Furthermore, the spatial posture of the slitting tool holder is adjusted according to the posture compensation instruction, and the key factor weight distribution of the process DOE model is iteratively updated based on the adjusted burr morphology data, including: Convert the posture compensation instruction into the compensation action of the slitting tool holder; Based on the maximum height and root curvature radius characteristics of the burr morphology after compensation, the weight proportional coefficient in the process DOE model is updated; When the fluctuation amount of the weight proportional coefficient updated by continuous iteration is lower than a preset threshold or reaches a maximum number of iterations, it is determined that the process DOE model has converged.

[0012] Furthermore, the weight proportional coefficient is updated using a momentum accelerated gradient descent method.

[0013] An intelligent tool setting control system, the system comprising: The real-time data acquisition module drives the slitting tool holder to cut the electrode and simultaneously collects the burr shape data and tool status parameters of the slitting edge; A key factor extraction module establishes a mapping relationship between burr generation and processing parameters based on a process DOE model, wherein the process DOE model is used to extract a set of key factors that affect slitting quality; A compensation instruction generation module generates a position compensation instruction for the slitting tool holder when the real-time monitoring value of the key factor set deviates from the optimization range of the process DOE model; A model weight adjustment module adjusts the spatial posture of the slitting tool holder according to the posture compensation instruction, and iteratively updates the key factor weight distribution of the process DOE model based on the adjusted burr morphology data; The tool setting control optimization module determines that the tool setting control closed-loop optimization is completed when the variance of the key factor weight distribution converges to a preset threshold.

[0014] Furthermore, the compensation instruction generating module includes: a deviation calculation unit, which calculates the deviation of the key control factor in the three-dimensional process space based on the real-time value of the key control factor; A compensation set generating unit generates an initial compensation parameter set by combining the burr feature contour distribution map and the tool health status data in the multi-dimensional correlation map; The feasibility screening unit performs feasibility screening on the initial compensation parameter set and outputs the posture compensation instruction that satisfies both the burr suppression target and the tool life protection.

[0015] The technical solution of the present invention can achieve the following technical effects: It effectively solves the defects of dynamic response hysteresis and insufficient compensation accuracy in traditional tool setting control. Through the three-dimensional dynamic compensation mechanism of real-time fusion of multi-source perception data, a closed-loop self-optimization control of the slitting tool holder posture-process parameters-burr suppression is constructed. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments recorded in the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0017] Figure 1 It is a flowchart of an intelligent tool setting control method; Figure 2 A flow chart for generating tool posture correction instructions; Figure 3 Schematic diagram of the process for generating a comprehensive dataset characterizing tool health; Figure 4 Schematic diagram of the process for adjusting tool posture and optimizing servo control parameters in real time; Figure 5 A schematic diagram of the process for generating an initial compensation parameter set; Figure 6 Schematic diagram of the process for adjusting the spatial posture of the slitting tool holder and updating the weight distribution of key factors. DETAILED DESCRIPTION

[0018] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments.

[0019] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the art to which this invention pertains. The terms used in this specification are for the purpose of describing specific embodiments only and are not intended to limit the invention. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0020] Example 1 like Figure 1As shown, the present invention provides an intelligent tool setting control method, the method comprising: S1: Drive the slitting tool holder to cut the electrode and simultaneously collect the burr shape data and tool status parameters of the slitting edge; Specifically, when the slitting tool holder performs the electrode cutting operation under the control of the servo drive system, a multi-source sensor collaborative acquisition mechanism is used to achieve full-dimensional monitoring of the processing process. At the moment the tool cuts into the material, a high-frequency linear array camera integrated on the side of the tool holder captures the slitting edge morphology with microsecond exposure. The image processing algorithm extracts the burr height, root crack depth, and waviness distortion characteristics in real time. Simultaneously, a multi-axis vibration sensor installed on the cutter head body and a thin film pressure sensor group on the tool clamping end respectively collect the tool radial vibration spectrum, axial pressure fluctuation curve, and contact stress gradient distribution. To determine the tool health status, a laser macro scanner is used to perform online three-dimensional reconstruction of the cutting edge profile. Combined with the acoustic emission sensor embedded in the tool handle, the cutting edge chip size, wear band width, and micro-crack extension trend are jointly monitored. All sensor data is transmitted to the edge computing node via a time-division multiplexing bus for timestamp alignment and feature parameter set packaging.

[0021] S2: The mapping relationship between burr generation and processing parameters is established based on the process DOE model. The process DOE model is used to extract the set of key factors affecting slitting quality. Specifically, by constructing a dynamic response model in a multi-dimensional process space, the complex coupling relationship between processing parameters and burr generation is systematically decoupled. The process DOE model uses historical processing data and real-time monitoring data to establish a nonlinear mapping network in multi-variable spaces such as slitting speed, tool pressure, and material ductility. It accurately identifies the core parameter combinations that are significantly sensitive to burr morphology under different working conditions. Through the integrated application of orthogonal experimental design and response surface method, a set of key factors that are simultaneously related to tool health status, material properties and equipment vibration modes (such as cutting edge stress concentration coefficient, plastic deformation energy threshold, etc.) are extracted, providing a quantifiable decision-making basis for subsequent dynamic compensation control and significantly reducing the optimization complexity of high-dimensional process parameter space.

[0022] S3: When the real-time monitoring value of the key factor set deviates from the optimization range of the process DOE model, a posture compensation instruction for the slitting tool holder is generated; Specifically, when the real-time processing status exceeds the process stability boundary, precise control instructions are quickly generated through the intelligent decision-making algorithm. By performing multi-dimensional threshold comparison of the real-time monitoring values ​​of key factors with the optimization interval preset by the process DOE model (such as tool stress margin, material plastic deformation threshold, etc.), the critical state of slitting quality degradation can be identified in time. Based on a comprehensive evaluation of the degree of deviation and the rate of change, the system dynamically calculates the displacement compensation, inclination angle correction and pressure adjustment gradient required for the slitting tool holder, and generates a compensation instruction set that includes the linkage between spatial posture adjustment parameters and process parameters, thereby achieving microsecond response under abnormal working conditions such as tool health decline and sudden change in material properties, and blocking the continuous deterioration of slitting quality defects.

[0023] S4: Adjust the spatial posture of the slitting tool holder according to the posture compensation instruction, and iteratively update the key factor weight distribution of the process DOE model based on the adjusted burr morphology data; Specifically, through the two-way interactive mechanism of dynamic posture adjustment and model parameter update, the process control system can realize adaptive optimization of time-varying processing conditions. After driving the slitting tool holder to complete the three-dimensional spatial posture correction according to the compensation instruction, the system reversely verifies the effectiveness of the compensation strategy based on the corrected slitting edge burr morphology characteristics (such as secondary forming burr height and root crack closure degree), and uses the reinforcement learning algorithm to gradient update the weight ratio of each key factor in the process DOE model. Through iterative optimization of the weight distribution, the model can autonomously enhance the sensitivity weight to dynamic disturbances such as tool degradation and material batch differences, while suppressing the interference effect of environmental noise, thus forming a technical closed loop of "compensation execution → quality verification → model evolution", and continuously improving the accuracy of tool control and process robustness.

[0024] S5: When the variance of the key factor weight distribution converges to a preset threshold, it is determined that the tool setting control closed-loop optimization is completed.

[0025] Specifically, during the iterative optimization process of the model, the dynamic changes of the weight distribution of key factors are continuously monitored. After each process parameter adjustment and burr morphology feedback are completed, the overall variance value of the current weight distribution is calculated and compared with the preset convergence threshold in real time. When the variance value is lower than the threshold for several consecutive iterations, the model is judged to have reached the steady-state convergence condition; if the number of iterations exceeds the preset maximum value and still does not meet the variance requirement, the tolerance mode is started to forcibly lock the weight distribution. After convergence judgment, the current optimal weight distribution parameters are solidified into the process knowledge base, and an optimization report containing compensation strategy effectiveness evaluation, tool life prediction and process window recommended values ​​is generated. At the same time, the closed-loop control signal is triggered to synchronize the slitting tool holder adjustment parameters to all machine equipment to ensure that subsequent processing processes are based on the latest model parameters.

[0026] Through the present invention, the defects of dynamic response hysteresis and insufficient compensation accuracy in traditional tool setting control are effectively solved. Through the three-dimensional dynamic compensation mechanism of real-time fusion of multi-source perception data, a closed-loop self-optimization control of the slitting tool holder posture-process parameters-burr suppression is constructed.

[0027] As a preferred embodiment of the above, Figure 2 As shown in the figure, the mapping relationship between burr generation and processing parameters is established based on the process DOE model, including: A10: Use the collected tool status parameters as coordinate axes to construct a three-dimensional process space, whose parameter range covers the entire tool life cycle and the fluctuation range of material properties; A20: Mapping real-time burr morphology data to the three-dimensional process space to generate a burr feature contour distribution map; A30: Mark tool wear gradient, edge chipping location and vibration energy level characteristics in the three-dimensional process space to form a multi-dimensional correlation map of process parameters-burr morphology-tool status.

[0028] Specifically, a dynamically expanded three-dimensional process space coordinate system is established with tool edge wear, radial pressure fluctuation amplitude, and material yield strength fluctuation value as orthogonal coordinate axes. The tool degradation trajectory throughout its life cycle (from the initial sharp state to the scrap threshold) is preloaded through historical processing data, and the characteristic fluctuation range of different batches of materials is preset based on the material library information. The octree space segmentation algorithm is used to realize the discrete modeling of the process parameter domain to ensure that each voxel unit accommodates the process parameter combination of a specific wear stage and material characteristics; the burr morphology data collected in real time is standardized and mapped to the three-dimensional process space through the bicubic spline interpolation algorithm to generate a burr feature contour distribution map with a continuous surface. For key quality indicators (such as burr height > 20μm), layered coloring technology is used to identify high-risk process areas, and the burr sensitivity index is calculated through a gradient operator to quantify the marginal impact of different process parameter combinations on slitting quality; a tool health status marking layer is embedded in the three-dimensional process space, and the tool wear inflection point is automatically identified using a point cloud clustering algorithm. The abnormal vibration energy level area is marked through the vibration spectrum feature extraction technology. Finally, the process parameters, burr distribution and tool status data are integrated to generate a multidimensional correlation map with spatiotemporal evolution characteristics. This supports the process model to quickly locate the coupling nodes of "high burr risk-tool sub-health-material abnormality", providing an explainable decision-making basis for the dynamic compensation strategy.

[0029] As a preferred embodiment of the above, Figure 3 As shown in the figure, the tool wear gradient, edge chipping location and vibration energy level characteristics are marked in the three-dimensional process space to form a multi-dimensional correlation map of process parameters-burr morphology-tool status, including: A31: A nonlinear mapping relationship is established based on historical data of tool wear gradient and slitting speed. Spatial clustering analysis is performed on the cutting edge chipping location parameters to generate a chipping-sensitive area distribution map. The high-frequency vibration risk area is marked by combining the frequency domain energy partitioning results of the vibration energy level characteristics. A32: The burr feature contour distribution map is stored in bins according to the material plastic deformation coefficient and dynamically associated with the tool health status data; A33: Construct a response surface model of the three-dimensional process vector including slitting speed, tool pressure, material coefficient, and the maximum burr height and root stress concentration coefficient.

[0030] Specifically, a long short-term memory network (LSTM) is used to perform nonlinear modeling on the temporal interaction between tool wear gradient and slitting speed. The weight matrix is ​​updated through gradient back propagation to establish a dynamic evolution equation of wear rate-processing intensity. After the coordinates of the chipping position on the cutting edge are transformed into three-dimensional spatial coordinates, the density clustering algorithm (DBSCAN) is used to identify the chipping space clustering area, and a chipping sensitive thermal map with probability density characteristics is generated. For the vibration energy level data, the frequency band energy ratio is extracted through wavelet packet decomposition, and the low-frequency stable domain, medium-frequency warning domain and high-frequency dangerous domain are divided. The vibration risk level is marked with different color temperatures in the process space. According to the plasticity of the material, the vibration risk level is determined by the material temperature. The deformation coefficient is discretized into bins (e.g., low ductility, medium ductility, and high ductility). Burr feature contour maps are stored hierarchically by material type. A dynamic time warping (DTW) algorithm is used to align tool wear trajectories with burr evolution curves in the time domain, constructing a correlation matrix for "material ductility-tool decay stage-burr growth rate." A three-dimensional process vector consisting of slitting speed, tool pressure, and material ductility coefficient is used as input variables to construct a dual-objective response surface for maximum burr height and root stress concentration coefficient. An adaptive sampling strategy is introduced to automatically increase the sampling point density in areas of sudden surface curvature changes, improving model prediction accuracy. The response surface model is integrated with the compensation control system in real time. A multi-objective particle swarm optimization algorithm is used to search for a Pareto-optimal process parameter combination that satisfies the dual constraints of burr suppression and stress control, and outputs the gradient descent direction of the three-dimensional posture compensation.

[0031] As a preferred embodiment of the above, Figure 4 As shown in FIG, when the real-time monitoring value of the key factor set deviates from the optimization range of the process DOE model, the position compensation instruction of the slitting tool holder is generated, including: S31: Based on the real-time value of the key control factor, calculate the deviation of the key control factor in the three-dimensional process space; S32: combining the burr feature contour distribution map in the multi-dimensional correlation map with the tool health status data to generate an initial compensation parameter set; S33: Perform feasibility screening on the initial compensation parameter set and output a posture compensation instruction that satisfies both the burr suppression target and the tool life protection.

[0032] Specifically, based on the three-dimensional process space coordinate system, the multi-dimensional deviation between the real-time monitoring value of the key factor and the model optimization interval is calculated by the weighted Euclidean distance algorithm. First, the key factors such as tool wear index, material plasticity coefficient, and vibration main frequency energy are normalized. According to the factor sensitivity weight distribution ratio output by the process DOE model, the total deviation scalar value is comprehensively calculated. At the same time, the sliding time window statistical method is used to analyze the dynamic change slope of the deviation, and the two types of abnormal modes, sudden deviation and gradual deviation, are identified, providing a pattern discrimination basis for the subsequent compensation strategy selection; different compensation decision paths are activated according to the deviation type. For sudden deviation, the adjacent process point set where the same type of burr is successfully suppressed is reversely searched in the burr feature contour distribution map, and its parameter combination such as slitting speed and tool inclination angle is extracted as the compensation benchmark. For gradual deviation, the tool health is combined with the tool health to determine the deviation. The health status data is collected and the pre-trained multi-objective optimization model is called to generate the Pareto frontier solution set. The parameter combinations output by the two types of paths are converted into the initial compensation parameter matrix containing the axial compensation amount of the tool holder, the radial inclination correction amount and the pressure adjustment gradient after coordinate transformation; a dynamic constraint condition library is established to verify the feasibility: the first is the mechanical execution constraint, which eliminates the parameter combinations that exceed the standard through the boundary collision detection algorithm; the second is the tool protection constraint, which uses the finite element simulation module to predict the tool stress concentration coefficient under each compensation scheme, and filters out the candidate schemes that exceed the fatigue limit; the third is the quality backtracking constraint, which adds the failed parameter combinations that cause secondary burrs in historical compensation cases to the blacklist library for matching and shielding, and finally uses the fuzzy comprehensive evaluation method to weightedly rank the burr suppression rate, tool loss rate and energy efficiency of the remaining candidate schemes, outputs the optimal posture compensation instruction set and sends it to the actuator.

[0033] As a preferred embodiment of the above, Figure 5 As shown, generate the initial compensation parameter set, including: B10: Based on the burr feature contour distribution map, select the direction with the maximum gradient change rate as the priority compensation axis; B20: Call the wear gradient in the tool health status data to calculate the compensation weight coefficient; B30: Generate an initial compensation parameter set on the priority compensation axis according to the compensation weight coefficient.

[0034] Specifically, the system analyzes the gradient trends of the burr feature contour distribution map and employs multi-directional gradient analysis to identify sensitive directions of process parameters. By grid-sampling dimensions such as slitting speed and tool pressure, the burr suppression efficiency along each axis is calculated. The core direction with the most significant adjustment effect (e.g., the axis where burr height drops sharply with increasing slitting speed) is identified as the priority compensation axis. Fine-tuning parameters in this direction maximizes burr suppression efficiency while mitigating overrun risks by integrating process boundary conditions. Based on real-time wear gradients (e.g., edge radius increment and wear zone width) derived from tool health data, the system dynamically balances burr suppression requirements with tool life objectives. When the tool is healthy (wear < lower threshold), the burr suppression weight is prioritized. When the tool enters a declining phase (remaining life < critical value), the compensation level is automatically reduced to delay failure. The weight coefficient is updated in real time with the tool degradation rate to ensure the sustainability of the compensation strategy. For the priority compensation axis, a base compensation amount is generated according to a dynamic weight ratio based on the deviation between the real-time parameter and the target value. Taking into account the material's plasticity (e.g., high-ductility materials require an enhanced pressure compensation gradient), material adaptation corrections are added. The resulting initial compensation parameter set is validated against multiple constraints: mechanical execution range, tool stress tolerance, and matching against a library of historical failure cases. High-risk solutions are eliminated, and the optimal instructions are output.

[0035] As a preferred embodiment of the above, the calculation formula of the compensation weight coefficient is: ; in, is the compensation weight coefficient, is the tool wear gradient, The maximum allowable tool wear threshold.

[0036] Specifically, The term establishes the ratio of the current wear rate (∇W) of the tool to the maximum allowable wear (W_max), which reflects the relative decline of the tool health. When the wear rate approaches the threshold, the ratio approaches 1. At this time, the compensation weight When it approaches 0, the system will automatically reduce the compensation amplitude to avoid aggravating tool damage. When the tool is in a healthy state (∇W is small), As the value approaches 1, the compensation strategy prioritizes burr suppression; as wear accelerates (∇W increases), As the value decreases, the control strategy gradually shifts to tool life protection, preventing secondary damage caused by overcompensation. This formula essentially establishes a negative feedback mechanism of "wear rate-compensation strength." For example, when edge chipping causes a sudden increase in ∇W, the formula will immediately reduce the compensation weight, suppressing the risk of stress concentration caused by aggressive adjustments while maintaining stable slitting quality.

[0037] As a preferred embodiment of the above, Figure 6 As shown in the figure, the spatial posture of the slitting tool holder is adjusted according to the posture compensation instruction, and the key factor weight distribution of the process DOE model is iteratively updated based on the adjusted burr morphology data, including: S41: converting the posture compensation instruction into the compensation action of the slitting tool holder; S42: based on the maximum height and root curvature radius characteristics of the burr morphology after compensation, updating the weight proportional coefficient in the process DOE model; S43: When the fluctuation amount of the weight proportional coefficient updated by continuous iterations is lower than a preset threshold or reaches the maximum number of iterations, it is determined that the process DOE model has converged.

[0038] Specifically, after receiving the posture compensation instruction, the slitting tool holder performs spatial posture adjustment through the multi-axis collaborative control mechanism, and the axial displacement compensation relies on the closed-loop linkage of the high-precision grating scale and the servo motor to achieve micron-level positioning calibration; the radial inclination correction is monitored by the gyroscope and combined with the hydraulic support system to complete the dynamic adjustment of the cutter disc flatness; the pressure gradient adjustment is based on the feedback of the material properties to match the blade contact pressure in real time. After the execution is completed, the laser scanner performs a full parameter review of the tool holder posture to ensure the accurate execution of the compensation action; then, the compensated burr morphology data is collected, and the characteristic parameters such as the maximum height, root curvature radius and crack density are extracted, and compared with the data before compensation for analysis: if the burr height is significantly reduced and the root morphology smoothness is improved, the compensation strategy is judged to be effective, and the model weights of key factors such as slitting speed and tool pressure are increased through the sliding window weighted algorithm; ... If the burr feature does not meet expectations or is accompanied by crack expansion, the influence weight of the material ductility factor is reduced, and the sensitivity distribution of the vibration energy level factor is enhanced. In this process, a dynamic smooth transition mechanism is adopted to avoid process oscillations caused by weight mutations. The standard deviation fluctuation of the weight ratio coefficient is continuously monitored during the model iteration. When the fluctuation amplitude is lower than the preset threshold for three consecutive times or reaches the maximum number of iterations, the convergence judgment mechanism is triggered: after successful convergence, the weight distribution is locked, a standard process parameter package is generated and synchronized to the production line equipment to form a global optimization; if convergence is not achieved, the historical optimal weight combination is enabled to maintain production, and the abnormal process window is marked for manual intervention diagnosis. Finally, a multi-dimensional evaluation report covering burr suppression rate, tool loss ratio and energy efficiency optimization is generated, completing the closed-loop optimization cycle from dynamic compensation to model self-learning, and realizing the continuous evolution of process stability and equipment economy.

[0039] As a preference of the above embodiment, the weight proportional coefficient is updated using a momentum accelerated gradient descent method.

[0040] Specifically, the momentum-accelerated gradient descent method is used to update the weight proportional coefficient in process parameter optimization, mainly based on its unique advantages in coping with complex industrial scenarios. In actual production, burr morphology data is easily disturbed by noise such as fluctuations in material properties and instantaneous vibration of the tool. The traditional gradient descent method is prone to weight oscillation due to single abnormal data, while the momentum mechanism effectively smooths random disturbances (such as burr height jumps caused by occasional vibrations) through the sliding average of historical gradients, making the update direction more in line with the actual process rules. At the same time, there are multi-dimensional coupling characteristics in the process parameter space (such as the interactive influence of slitting speed and tool pressure on burrs). The inertia effect of the momentum method can accelerate the crossing of local optimal traps. For example, in the middle stage of tool wear, it can quickly lock the balance point that takes into account burr suppression and life extension. In the face of dynamic scenarios (such as material batch switching and progressive tool wear), this method can respond to new data features in a timely manner by retaining historical gradient memory, while avoiding excessive forgetting of verified effective parameter combinations.

[0041] Example 2 Based on the same inventive concept as the intelligent tool setting control method in the aforementioned embodiment, the present invention further provides an intelligent tool setting control system, the system comprising: The real-time data acquisition module drives the slitting tool holder to cut the electrode and simultaneously collects the burr shape data and tool status parameters of the slitting edge; The key factor extraction module establishes the mapping relationship between burr generation and processing parameters based on the process DOE model. The process DOE model is used to extract the key factor set that affects the slitting quality. The compensation instruction generation module generates the position compensation instruction of the slitting tool holder when the real-time monitoring value of the key factor set deviates from the optimization range of the process DOE model; The model weight adjustment module adjusts the spatial posture of the slitting tool holder according to the posture compensation instruction, and iteratively updates the key factor weight distribution of the process DOE model based on the adjusted burr morphology data; The tool setting control optimization module determines that the tool setting control closed-loop optimization is completed when the variance of the key factor weight distribution converges to a preset threshold.

[0042] The above-mentioned control system in the present invention can effectively implement the intelligent tool setting control method, and the technical effects that can be achieved are as described in the above-mentioned embodiments, which will not be repeated here.

[0043] As a preferred embodiment of the above, the compensation instruction generation module includes: Deviation calculation unit, which calculates the deviation of key control factors in the three-dimensional process space based on the real-time values ​​of key control factors; The compensation set generation unit generates an initial compensation parameter set by combining the burr feature contour distribution map in the multi-dimensional correlation map and the tool health status data; The feasibility screening unit performs feasibility screening on the initial compensation parameter set and outputs posture compensation instructions that simultaneously meet the burr suppression goals and tool life protection.

[0044] Similarly, the above-mentioned optimization schemes for the system can also respectively achieve the corresponding optimization effects of the method in Example 1, which will not be repeated here.

[0045] Although the present application has been described with reference to specific features and embodiments thereof, it is apparent that various modifications and combinations thereof may be made without departing from the spirit and scope of the present application. Accordingly, this specification and the drawings are merely illustrative of the present application as defined herein, and are intended to cover any and all modifications, variations, combinations or equivalents within the scope of the present application. Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the present application and its equivalents, the present application is intended to include these modifications and variations.

Claims

1. An intelligent tool setting control method, characterized in that: include: Drive the slitting tool holder to cut the electrode, and simultaneously collect the burr shape data and tool status parameters of the slitting edge; Establishing a mapping relationship between burr generation and processing parameters based on a process DOE model, wherein the process DOE model is used to extract a set of key factors affecting slitting quality; When the real-time monitoring value of the key factor set deviates from the optimization range of the process DOE model, a posture compensation instruction of the slitting tool holder is generated; Adjusting the spatial posture of the slitting tool holder according to the posture compensation instruction, and iteratively updating the key factor weight distribution of the process DOE model based on the adjusted burr morphology data; When the variance of the key factor weight distribution converges to a preset threshold, the tool setting control closed-loop optimization is determined to be completed.

2. The intelligent tool setting control method according to claim 1, characterized in that: The mapping relationship between burr generation and machining parameters is established based on the process DOE model, including: The tool state parameters collected are used as coordinate axes to construct a three-dimensional process space, whose parameter range covers the entire tool life cycle and the fluctuation range of material properties; Mapping the real-time burr morphology data to a three-dimensional process space to generate a burr feature contour distribution map; The tool wear gradient, cutting edge chipping position and vibration energy level characteristics are marked in the three-dimensional process space to form a multi-dimensional correlation map of process parameters-burr morphology-tool status.

3. The intelligent tool setting control method according to claim 2, characterized in that: Marking the tool wear gradient, edge chipping location, and vibration energy level characteristics in the three-dimensional process space to form a multi-dimensional correlation map of process parameters, burr morphology, and tool status, including: A nonlinear mapping relationship is established based on the historical data of the tool wear gradient and the slitting speed. A spatial cluster analysis is performed on the cutting edge chipping position parameters to generate a chipping sensitive area distribution map. The high-frequency vibration risk domain is marked in combination with the frequency domain energy partition result of the vibration energy level characteristics. The burr feature contour distribution map is stored in bins according to the material plastic deformation coefficient, and is dynamically associated with the tool health status data; A response surface model including slitting speed, tool pressure, material coefficient, three-dimensional process vector, maximum burr height and root stress concentration coefficient was constructed.

4. The intelligent tool setting control method according to claim 1, characterized in that: When the real-time monitoring value of the key factor set deviates from the optimization range of the process DOE model, a posture compensation instruction for the slitting tool holder is generated, including: Calculating the deviation of the key control factor in the three-dimensional process space based on the real-time value of the key control factor; generating an initial compensation parameter set by combining the burr feature contour distribution map in the multi-dimensional correlation map with the tool health status data; The initial compensation parameter set is screened for feasibility, and the posture compensation instruction that satisfies both the burr suppression goal and the tool life protection is output.

5. The intelligent tool setting control method according to claim 4, characterized in that: Generating the initial compensation parameter set includes: According to the burr feature contour distribution map, the direction with the maximum gradient change rate is selected as the priority compensation axis; Calling the wear gradient in the tool health status data to calculate the compensation weight coefficient; The initial compensation parameter set is generated on the priority compensation axis according to the compensation weight coefficient.

6. The intelligent tool setting control method according to claim 5, characterized in that: The calculation formula of the compensation weight coefficient is: ； in, is the compensation weight coefficient, is the tool wear gradient, The maximum allowable tool wear threshold.

7. The intelligent tool setting control method according to claim 1, characterized in that: The spatial posture of the slitting tool holder is adjusted according to the posture compensation instruction, and the key factor weight distribution of the process DOE model is iteratively updated based on the adjusted burr morphology data, including: Convert the posture compensation instruction into the compensation action of the slitting tool holder; Based on the maximum height and root curvature radius characteristics of the burr morphology after compensation, the weight proportional coefficient in the process DOE model is updated; When the fluctuation amount of the weight proportional coefficient updated by continuous iteration is lower than a preset threshold or reaches a maximum number of iterations, it is determined that the process DOE model has converged.

8. The intelligent tool setting control method according to claim 1, characterized in that: The weight proportional coefficient is updated using the momentum accelerated gradient descent method.

9. An intelligent tool setting control system, characterized in that: The system comprises: The real-time data acquisition module drives the slitting tool holder to cut the electrode and simultaneously collects the burr shape data and tool status parameters of the slitting edge; A key factor extraction module establishes a mapping relationship between burr generation and processing parameters based on a process DOE model, wherein the process DOE model is used to extract a set of key factors that affect slitting quality; A compensation instruction generation module generates a position compensation instruction for the slitting tool holder when the real-time monitoring value of the key factor set deviates from the optimization range of the process DOE model; A model weight adjustment module adjusts the spatial posture of the slitting tool holder according to the posture compensation instruction, and iteratively updates the key factor weight distribution of the process DOE model based on the adjusted burr morphology data; The tool setting control optimization module determines that the tool setting control closed-loop optimization is completed when the variance of the key factor weight distribution converges to a preset threshold.

10. The intelligent tool setting control system according to claim 9, characterized in that: The compensation instruction generation module includes: a deviation calculation unit, which calculates the deviation of the key control factor in the three-dimensional process space based on the real-time value of the key control factor; A compensation set generating unit generates an initial compensation parameter set by combining the burr feature contour distribution map and the tool health status data in the multi-dimensional correlation map; The feasibility screening unit performs feasibility screening on the initial compensation parameter set and outputs the posture compensation instruction that satisfies both the burr suppression target and the tool life protection.

Images