Cutting control method, system and device for machining of metal parts

Abstract

This invention discloses a cutting control method, system, and equipment for metal parts processing, relating to the field of intelligent manufacturing technology. The method includes: simultaneously and in real-time acquiring in-situ sensing data from at least two physical domains—vibration, acoustic emission, heat, and microstructure—during the cutting process; driving and updating a process-level digital twin based on this data in real-time to predict future deformation and defect risks; performing multi-objective optimization based on the prediction results to generate an optimal composite intervention strategy; controlling the cutting equipment to execute the strategy and returning to the sensing step, forming a closed-loop adaptive control. This invention solves the technical problem of existing metal cutting methods being unable to monitor and adjust residual stress and deformation during the cutting process in real-time, leading to poor cutting accuracy and stability. It achieves the technical effect of improving the cutting accuracy and stability of metal parts under complex stress states by predicting and actively compensating for deformation defects caused by stress release in real-time based on multi-domain sensing and digital twins.

Description

Technical Field

[0001] This invention relates to the field of intelligent manufacturing technology, and specifically to cutting control methods, systems and equipment for processing metal parts. Background Technology

[0002] Residual stress poses a significant challenge in the machining of metal parts, especially large, complex, or heat-treated components. Traditional cutting processes, such as laser, waterjet, or wire EDM, can cause a sudden redistribution of internal residual stress when severing material connections, leading to unpredictable instantaneous deformation, dimensional deviations, and even cracking of the workpiece. This problem is particularly prominent in aerospace and mold manufacturing, severely restricting the yield rate and material utilization of high-precision, high-performance components. Existing technologies often focus on offline simulation to optimize process parameters or rely on single sensors for local compensation, which struggles to address the real-time and nonlinear effects of the dynamic release of residual stress during machining, resulting in insufficient machining quality and stability. Summary of the Invention

[0003] This application provides a cutting control method, system, and equipment for processing metal parts, which solves the technical problem that existing metal cutting methods cannot monitor and adjust residual stress and deformation generated during the cutting process in real time, resulting in poor cutting accuracy and stability.

[0004] The first aspect of this application provides a cutting control method for processing metal parts. The method includes: synchronously and in real-time acquiring in-situ sensing data from at least two physical domains during the movement of the cutting point, the physical domains including a vibration domain, an acoustic emission domain, a thermal domain, and a microscopic morphology domain; based on the in-situ sensing data, driving and updating a process-level digital twin of the metal parts in real time, solving the coupled field model of the process-level digital twin, and predicting the component deformation trend and potential defect initiation risk of the current cutting action within a future time window, wherein the process-level digital twin couples a structural mechanical field, a thermal conduction field, and a material phase transition field; based on the component deformation trend and the potential defect initiation risk, performing multi-objective optimization to generate an optimal composite intervention strategy for the next time step; controlling the cutting equipment to execute the optimal composite intervention strategy, and returning to the acquisition of in-situ sensing data after execution for closed-loop adaptive control.

[0005] A second aspect of this application provides a cutting control system for metal parts processing. The system includes: a data synchronization acquisition module for synchronously and in real-time acquiring in-situ sensing data from at least two physical domains during the movement of the cutting point, the physical domains including a vibration domain, an acoustic emission domain, a thermal domain, and a microscopic morphology domain; a cutting risk prediction module for driving and updating a process-level digital twin of the metal parts in real-time based on the in-situ sensing data, solving the coupled field model of the process-level digital twin, and predicting the component deformation trend and potential defect initiation risk within a future time window, wherein the process-level digital twin couples a structural mechanical field, a thermal conduction field, and a material phase transition field; a multi-objective optimization solution module for performing multi-objective optimization based on the component deformation trend and the potential defect initiation risk, and determining the optimal composite intervention strategy for the next time step; and a cutting execution feedback module for controlling the cutting equipment to execute the optimal composite intervention strategy, and returning to the acquisition of in-situ sensing data after execution for closed-loop adaptive control.

[0006] A third aspect of this application provides an electronic device comprising: a processor coupled to a memory for storing a program that, when executed by the processor, causes the system to perform the method described in any of the first aspects.

[0007] One or more technical solutions provided in this application have at least the following technical effects or advantages: This application provides a cutting control method, system, and equipment for metal parts processing, relating to the field of intelligent manufacturing technology. By constructing an adaptive control loop of in-situ multi-domain perception, digital twin dynamic prediction, multi-objective optimization decision-making, and closed-loop collaborative intervention, it senses the material stress state in real time and drives the coupled field twin to predict ahead. It generates and executes a composite intervention strategy that integrates energy, trajectory, and medium online, achieving proactive suppression of deformation and defects induced by residual stress. This solves the technical problem of existing metal cutting methods being unable to monitor and adjust residual stress and deformation generated during the cutting process in real time, leading to poor cutting accuracy and stability. It achieves closed-loop adaptive control based on multi-domain perception and digital twin, predicting and proactively compensating for deformation and defects caused by stress release in real time, thus improving the cutting accuracy and stability of metal parts under complex stress states. Attached Figure Description

[0008] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0009] Figure 1 A schematic flowchart of a cutting control method for processing metal parts provided in an embodiment of this application; Figure 2 This is a schematic diagram of a cutting control system for metal parts processing provided in an embodiment of this application; Figure 3 This application provides a schematic diagram of the structure of an electronic device.

[0010] Explanation of reference numerals in the attached figures: 11 data synchronization acquisition module, 12 cutting risk prediction module, 13 protection multi-objective optimization solution module, 14 cutting execution feedback module, 300 electronic device, 301 memory, 302 processor, 303 communication interface, 304 bus architecture. Detailed Implementation

[0011] This application provides a cutting control method, system, and equipment for processing metal parts, which solves the technical problem that existing metal cutting methods cannot monitor and adjust residual stress and deformation generated during the cutting process in real time, resulting in poor cutting accuracy and stability.

[0012] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0013] It should be noted that the terms "first," "second," etc., in the specification and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or server that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or modules not explicitly listed or inherent to such processes, methods, products, or devices.

[0014] Example 1, as Figure 1 As shown, this application provides a cutting control method for machining metal parts, the method comprising: P10: During the movement of the cutting point, in-situ sensing data of at least two physical domains are collected synchronously in real time. The physical domains include the vibration domain, acoustic emission domain, thermal domain, and micromorphological domain.

[0015] Furthermore, step P10 in this embodiment of the application also includes: P11: Real-time acquisition of microscopic vibration signals in the processed area behind the cutting point using a laser Doppler vibrometer to obtain in-situ sensing data in the vibration domain; P12: Real-time acquisition of elastic wave signals excited by microscopic material fractures during the cutting process using an acoustic emission sensor array to obtain in-situ sensing data in the acoustic emission domain; P13: Real-time acquisition of dynamic temperature field signals including the cutting seam and heat-affected zone using a high-speed infrared thermal imager to obtain in-situ sensing data in the thermal domain; P14: Real-time acquisition of microscopic geometric morphology signals of the cutting seam sidewall using an optical coherence tomography module to obtain in-situ sensing data in the micromorphology domain.

[0016] It should be understood that, during the movement of the cutting point, in order to capture in real time the multi-dimensional changes in the internal state of the workpiece caused by the release of residual stress, this application simultaneously and in parallel implemented high-frequency in-situ sensing across multiple physical domains. The core of this sensing system lies in ensuring that the data streams acquired from different physical domains are strictly synchronized in timestamps and precisely registered in spatial coordinates, thereby providing input for the subsequent construction of a spatiotemporally consistent dynamic profile of the process. Specifically, the sensing implementation is achieved by integrating a set of high-sensitivity sensor arrays. This sensor array is deployed near the cutting head and moves in tandem with it to ensure that the observation field of view is always focused on the dynamic process area centered on the cutting point.

[0017] At the vibration domain sensing level, a laser Doppler vibrometer is used. Its laser probe aims at and tracks a specific surface point in the machined area behind the cutting point at a fixed offset distance. Based on the principle of optical heterodyne interferometry, this instrument can non-contactly measure the instantaneous vibration velocity of the measured point along the laser beam direction. During the cutting process, the uneven release of residual stress within the workpiece triggers continuous changes in local stiffness, which directly modulate the vibration response characteristics of that point. The vibrometer continuously records the velocity time-domain signal of this point at a sampling rate of at least 50 kHz and converts it into frequency-domain data in real time through a built-in signal processing unit. Feature vectors, including first to several orders of key resonant frequencies and their corresponding amplitudes, are extracted, forming a vibration domain in-situ sensing data stream reflecting the evolution of the structure's dynamic characteristics. This data is a key indicator for subsequently determining whether the overall behavior of the workpiece has drifted.

[0018] At the acoustic emission sensing level, a ring array consisting of four or more piezoelectric ceramic acoustic emission sensors is employed, arranged centered on the cutting path on a specific coupling agent on the workpiece surface. When the material undergoes micro-yielding, dislocation slip, or microcrack nucleation under the combined effects of cutting thermal load and residual stress, transient elastic waves are released. The acoustic emission sensor array synchronously acquires these high-frequency signals at a sampling rate of no less than one megahertz. The acquired raw waveform data is transmitted to an acoustic emission analyzer, which first performs threshold detection and parameter extraction on the signal of each channel, obtaining characteristic parameters including impact count, energy, amplitude, rise time, and duration. Furthermore, by analyzing the time difference of the same event signal arriving at different sensors, combined with a known wave velocity model, the two-dimensional or three-dimensional coordinates of the acoustic emission source in the workpiece can be calculated in real time, achieving preliminary localization of the damage event. The processed acoustic emission event characteristic parameters and the localization information together constitute the in-situ sensing data in the acoustic emission domain, which is the most direct and sensitive criterion for predicting the risk of materials entering the nonlinear damage stage and defect initiation.

[0019] At the thermal sensing level, a high-speed infrared thermal imager is employed, its spectral response range corresponding to the main infrared radiation bands of the tested metal material at typical processing temperatures. The imager's optical axis is calibrated to a fixed tilt angle, aligning with the area containing the cutting kerf and its two heat-affected zones, ensuring its field of view completely covers the spatial range where the main thermal processes occur. The imager continuously captures images at a frame rate of at least 100 frames per second, generating a series of temporally continuous spatial temperature distribution images. Each pixel value in each frame is converted into the absolute temperature value of the workpiece surface corresponding to that pixel using a pre-calibrated radiation model and temperature calibration curve. This allows for the acquisition of the complete two-dimensional temperature field evolution over time of the molten pool at the cutting kerf front, the solidified cut, and the heat-affected zone, including characteristic quantities such as temperature gradient, rate of change of high-temperature area, and the movement speed of specific isotherms. This dynamic temperature field signal is the core of the in-situ thermal sensing data; it quantitatively describes the dynamic balance between the cutting heat input and the workpiece heat dissipation, serving as a key boundary condition driving the thermo-mechanical coupling simulation.

[0020] At the micro-morphology domain sensing level, an integrated optical coherence tomography module based on a swept-frequency light source is employed. The module's sampling probe incorporates a coaxial illumination interferometric optics system, with its low-coherence beam focused at an approximately perpendicular angle onto the sidewall surface of the cut. The beam performs an axial scan of the surface, and by detecting the interference signal of the backscattered light, a depth-resolved reflectivity profile along the beam direction can be reconstructed. A lateral scan yields a three-dimensional image of the cut sidewall surface and its subsurface microstructure. During the cutting process, the module repeatedly performs three-dimensional scans at specific time intervals, acquiring real-time micro-geometric morphology signals of the cut surface. From this three-dimensional data, key morphological parameters such as the arithmetic mean of the sidewall surface roughness, the height of the profile peaks and valleys, and the presence of microcracks wider than one micrometer can be quantitatively extracted. These parameters provide direct evidence for assessing the cut surface quality and monitoring whether the processed area has developed micro-tears or deformation wrinkles due to stress release, constituting in-situ sensing data in the micro-morphology domain.

[0021] Ultimately, the sensing systems of the four physical domains are coordinated by a central synchronous controller, ensuring that the data acquisition trigger pulses of the laser Doppler vibrometer, the clock of the acoustic emission acquisition system, the frame exposure signal of the high-speed infrared thermal imager, and the scanning trigger signal of the optical coherence tomography module are all phase-locked to the same high-precision time base. All acquired vibration spectrum feature vectors, acoustic emission event parameter lists, two-dimensional temperature matrix sequences, and three-dimensional topographic point cloud data streams are appended with a unified timestamp and the real-time position coordinates of the cutting head, and transmitted to the data processing unit for alignment and fusion, thereby generating a time-synchronized multimodal in-situ sensing dataset, providing real-time updated input data for the process-level digital twin.

[0022] P20: Based on the in-situ sensing data, the process-level digital twin of the metal parts is driven and updated in real time, the coupled field model of the process-level digital twin is solved, and the deformation trend of the component and the risk of potential defect initiation in the future time window are predicted. The process-level digital twin is coupled with the structural mechanical field, the heat conduction field and the material phase transition field.

[0023] Furthermore, in this embodiment, step P20 further includes driving and updating the process-level digital twin of metal parts in real time: P21: Based on the geometric, material, and constraint information of the metal parts, construct a basic digital twin containing an initial residual stress field; P22: In the basic digital twin, configure a coupled field model that couples a heat source model, a heat transfer model, a microstructure evolution model, and a stress-strain calculation model, wherein the heat source model is used to map the control parameters of the cutting energy source to the heat flux density distribution parameters applied to the workpiece surface; P23: Based on the in-situ sensing data, update the basic digital twin in real time, including: P23-1: Based on the in-situ sensing data of the thermal domain and the current energy source control parameters, the heat flux density distribution parameters of the heat source model are inverted in real time to calibrate the energy input; P23-2: Based on the calibrated energy input, combined with the in-situ sensing data of the thermal domain, the thermophysical parameters of the heat transfer model are inverted and calibrated in real time to update the heat conduction field; P23-3: Through the in-situ sensing data of the acoustic emission domain and the in-situ sensing data of the micromorphology domain, material damage characteristics are identified, and the tissue evolution model is corrected in real time to update the material phase transition field; P23-4: Based on the in-situ sensing data of the vibration domain and the updated heat conduction field and the material phase transition field, the stress-strain calculation model is corrected in real time to update the structural mechanical field.

[0024] Optionally, based on the multimodal in-situ sensing dataset obtained in the aforementioned steps, a high-fidelity process-grade digital twin is dynamically driven and updated. This process-grade digital twin is not a static model, but a dynamic virtual entity capable of evolving synchronously with the physical workpiece and undergoing ultra-real-time simulation. By solving the coupled multiphysics model within it, the deformation trend and potential defect initiation risk of the component within a predetermined time window can be predicted under the assumption that the current cutting action remains unchanged. This process-grade digital twin deeply integrates structural mechanical fields, thermal conduction fields, and material phase transition fields to accurately characterize the complex interactions of heat, force, and structure during the cutting process.

[0025] In practical implementation, the initial construction of a process-level digital twin must first be completed. For example, a basic digital twin is constructed based on the computer-aided design geometric model of the target metal part, the basic physical property parameter library corresponding to the material grade, and the constraint information for fixture positioning. This basic twin is spatially discretized using the finite element method, with the mesh refined in the region near the cutting path to ensure computational accuracy. Crucially, the initial residual stress field distribution data inside the workpiece, obtained in advance through various measurement methods such as X-ray diffraction or ultrasonic methods, must be used as the initial stress condition for each integration point of the finite element mesh-discrete basic twin. This ensures that a high-fidelity digital copy reflecting the true initial mechanical state of the workpiece is established at the start of the simulation.

[0026] Next, a coupled field model, which couples multiple sub-models, is configured within the basic digital twin framework. This coupled field model contains at least four key sub-models, which are bidirectionally or unidirectionally coupled through shared field variables to form a closed-loop solution system. The first is the heat source model, which maps real-time control parameters of the cutting energy source, such as laser power, spot size, and moving speed, to the spatiotemporal heat flux density applied to the surface units of the workpiece cutting path through a physical formula based on Gaussian distribution. The second is the heat transfer model, constructed based on the finite element equations of transient nonlinear heat conduction, used to calculate the temperature field distribution and evolution inside the workpiece. Its solution depends on the material's thermal conductivity, specific heat capacity, density, and other physical properties that change with temperature. The third is the microstructure evolution model, which is based on continuous cooling transformation kinetic curves and uses models such as Leblond or Koistinen-Marburger to predict the phase transformation process of the material's microstructure based on the temperature history of the unit, outputting the volume fractions of each phase, such as austenite and martensite, and updating the material property parameters of the unit accordingly. The fourth is the stress-strain calculation model, which adopts a thermo-elastic-plastic constitutive relation. Under the quasi-static assumption, it solves the equilibrium equations using an incremental iterative method to calculate the stress-strain field inside the workpiece and the resulting deformation. This model needs to consider both the thermal strain caused by the temperature field and the phase transformation strain caused by the phase transformation. These sub-models are solved sequentially within a unified time step and coupled through field variable transfer. For example, the heat input generated by the heat source model is transferred as a load to the heat transfer model. The temperature field results calculated by the heat transfer model are transferred to the microstructure evolution model and the stress-strain calculation model. The phase transformation strain and material property changes calculated by the microstructure evolution model are then fed back as inputs to the stress-strain calculation model.

[0027] Subsequently, the real-time driving and updating phase begins. Utilizing the in-situ sensing data acquired in the previous steps, key components of the coupled field model are calibrated and corrected online within each control cycle to ensure the state of the digital twin remains synchronized with the physical workpiece. The specific update process includes four parallel sub-steps. First, based on real-time temperature field data acquired by a high-speed infrared thermal imager, particularly the highest temperature value at the center of the molten pool at the cutting kerf edge and the geometric dimensions of the molten pool, combined with the current power and spot diameter parameters of the cutting laser, an inverse problem optimization model is constructed. This model uses the energy absorption rate and effective Gaussian radius from the heat source model as the parameters to be inverted, and the least squares error between the simulated molten pool characteristic dimensions and temperature and the measured values ​​as the objective function. It employs algorithms such as the Levenberg-Marquardt algorithm for solving the problem, dynamically calibrating the energy input terms in the model to more accurately reflect the actual heat input.

[0028] Next, based on the inverted heat source model and in-situ sensing data of the thermal domain, the thermophysical parameters in the heat transfer model are retrieved in real time. Specifically, this process involves analyzing temperature changes during heat conduction and calibrating thermophysical parameters such as thermal conductivity and specific heat capacity of the metallic material in real time. The steps include acquiring temperature data in real time during the cutting process and comparing it with the output of the heat transfer model. By comparing the actual measured values ​​with the predicted values ​​of the temperature field, the thermophysical parameters in the heat transfer model are retrieved. The heat transfer model is then adjusted using optimization algorithms or inversion techniques to ensure a more accurate simulation of heat transfer within the metallic component, thereby optimizing the calculation results of the heat conduction field. Through this process, the temperature distribution of the heat-affected zone during cutting and its impact on material properties can be predicted more accurately.

[0029] Next, in-situ sensing data from both the acoustic emission domain and the microstructure domain are used to identify material damage characteristics and to make real-time corrections to the tissue evolution model. Acoustic emission data can provide early signals of material cracks and damage during the cutting process, while microstructure data can help accurately capture microscopic defects or morphological changes on the material surface. First, feature extraction is performed on the waveform signals acquired by the acoustic emission sensor array, including the energy, absolute energy, ring count, and peak frequency of each event. When the cumulative count of high-energy events exceeds a threshold set based on historical data statistics within a time window, it is determined that the material damage activity in that area has significantly increased. Simultaneously, the three-dimensional point cloud data of microstructure provided by optical coherence tomography is processed to detect microcracks with widths greater than a preset threshold, such as one micrometer. When both acoustic emission features and microstructure observations indicate damage, a correction to the tissue evolution model is triggered. Specifically, the Ms point used to predict the onset of martensitic phase transformation will be adjusted, or the exponential parameters in the Avrami equation describing phase transformation kinetics will be modified, so that the microstructure transformation predicted by the model under the same thermal history is more likely to generate brittle phases, thereby updating the prediction of the material phase transformation field and making it closer to the actual damage state of the material under complex thermal loads.

[0030] Finally, the structural dynamic characteristics provided by the vibration domain sensing data are used to correct the structural mechanical field calculations. The real-time vibration velocity time-domain signal obtained by the laser Doppler vibration meter is processed by a Fast Fourier Transform to obtain the real-time frequency response function of the workpiece. From this frequency response function, the top three dominant natural frequencies during the cutting process and their corresponding damping ratios are identified. The natural frequencies calculated by modal analysis of the digital twin in its current state (after updating the temperature and material property fields as described above) are compared with the measured natural frequencies. If the error of a certain frequency exceeds the allowable range, a sensitivity-based inversion algorithm is used to fine-tune the equivalent elastic modulus of the corresponding region in the twin model, or the boundary stiffness parameters of key fixture connections, so that the dynamic characteristics of the stress-strain calculation model are consistent with the dynamic characteristics of the physical workpiece in the measured vibration data. This updates the structural mechanical field, making the predicted deformation trend more reliable.

[0031] Through the iterative execution of the above four sub-steps in each control cycle, the process-level digital twin is continuously driven and updated by in-situ sensing data, ensuring that its state is kept as synchronized as possible with the actual state of the physical workpiece during processing. This process ensures precise control of the cutting process and timely identification of potential defects and deformation risks, providing reliable data support for subsequent process adjustments.

[0032] Furthermore, by solving the coupled field model of the process-level digital twin and predicting the component deformation trend and potential defect initiation risk within a future time window, step P20 of this embodiment also includes: P24: Based on the coupled field model after real-time inversion calibration and real-time feedback correction using in-situ sensing data, perform advanced numerical simulation and record the advanced simulation data; P25: Based on the advanced simulation data, extract the full-field displacement data of the metal components within a future time window, and calculate the root mean square value of the normal deviation of the full-field displacement data relative to the ideal geometric profile to obtain the deformation trend of the component; P26: Based on the advanced simulation data, extract the stress intensity factor of the stress concentration region in the metal components within a future time window, and calculate the safety margin of the stress intensity factor relative to the material fracture toughness, using the reciprocal of the safety margin as the potential defect initiation risk.

[0033] Specifically, based on the high-fidelity coupled field model after real-time inversion calibration and real-time feedback correction of in-situ sensing data, advanced numerical simulation is performed, and key predictive indicators for decision-making, namely component deformation trends and potential defect initiation risks, are quantitatively extracted as direct inputs for subsequent optimization decisions.

[0034] Specifically, firstly, advanced numerical simulations are performed based on the coupled field model, which has undergone real-time inversion calibration and real-time feedback correction using in-situ sensing data. The key to this process is that the model, after inversion and correction of the real-time sensing data, can predict and simulate the cutting process within a future time window. The purpose of advanced simulation is to simulate the physical changes that the metal components may undergo during the subsequent cutting process, especially changes in deformation and stress distribution. During the simulation, the updated coupled field model needs to be applied to finite element analysis (FEA) or other numerical simulation methods to predict the cutting process. This simulation process not only calculates the temperature, stress, and displacement fields but also considers the nonlinear behavior of the material, phase transition effects, and thermo-mechanical coupling effects. All data generated during the simulation, including temperature changes, stress changes, and deformation, will be recorded as advanced simulation data for subsequent analysis and optimization.

[0035] Subsequently, based on advanced simulation data, the full-field displacement data of the metal components within a future time window is extracted. Full-field displacement data refers to the displacement of each node of the metal component during the cutting process. This data is obtained through numerical simulation and can describe the material deformation over a future time period. Using this displacement data, the root mean square (RMS) value of the normal deviation from the ideal geometric profile can be calculated. The ideal geometric profile refers to the geometric shape that the metal component should achieve according to design or processing requirements, while the normal deviation refers to the difference between the actual deformation and this ideal shape. Calculating the RMS value provides a quantitative indicator to measure the deformation trend of the component during cutting. A larger RMS value indicates more severe deformation, thus providing guidance for subsequent cutting path adjustments and process optimization. Through this process, the deformation trend of the component can be predicted in real time, providing data support for avoiding the production of defective products.

[0036] Simultaneously, the potential defect initiation risk is extracted and calculated from the same set of advanced simulation data. First, based on the simulation data, high stress concentration areas that persist or newly appear in the metal components within future time windows are identified. For example, a threshold, such as 80% of the material's yield strength, can be set to filter out high-risk elements with equivalent stress exceeding this threshold. For these high-risk elements, their stress intensity factors are further calculated. The stress intensity factor is an important parameter describing the stress state near the crack tip in a material and is a key indicator for predicting crack propagation. For example, for predicted potential Type I cracks, the stress intensity factor KI of these elements or element fronts can be calculated using the interactive integration method based on finite element results. Next, the fracture toughness KIC of the metal material at the current predicted temperature and microstructure is retrieved from the material database. For each location where the stress intensity factor is calculated, at each future time, its safety margin, i.e., the ratio of fracture toughness KIC to stress intensity factor KI, is calculated. A safety margin greater than 1 indicates safety, while a margin less than 1 indicates a risk of unstable propagation. To obtain a dimensionless index positively correlated with risk, the reciprocal of the safety margin, KI / KIC, is used as the defect initiation risk index for that location at that time. Finally, the maximum value of this risk index for all high-risk locations across all future times, or a composite value obtained by integrating or averaging it over future time windows, is taken as the quantitative output of the potential defect initiation risk. A higher risk index value indicates a higher risk of potential defects. This data can be used to guide the cutting process in identifying areas where defects may occur, allowing for appropriate interventions to prevent crack propagation and component failure. The advantage of this method is that it enables real-time adjustment of the cutting strategy, ensuring workpiece processing quality, reducing the risk of deformation, cracks, or dimensional deviations caused by residual stress, and improving the accuracy and reliability of process control.

[0037] Furthermore, based on the advanced simulation data, the stress intensity factor of the stress concentration region in the metal component within the future time window is extracted to obtain the potential defect initiation risk. Step P20 in this embodiment further includes: P27: During the advanced simulation process, regions in the metal components where the principal stress exceeds the material's yield strength are identified and recorded as stress concentration regions; P28: For each stress concentration region, based on the advanced simulation data, the Type I, Type II, and Type III stress intensity factors are calculated in different preset propagation directions; P29: The Type I, Type II, and Type III stress intensity factors are combined to obtain a composite stress intensity factor for the stress concentration region, and the ratio of the composite stress intensity factor to the material's dynamic fracture toughness is calculated as a quantitative indicator of the potential defect initiation risk.

[0038] In one possible embodiment of this application, to further accurately predict potential defects that may occur in metal parts during the cutting process, particularly the risk of crack initiation in stress concentration areas, the process of identifying risk areas and calculating evaluation indicators can be further refined. This process aims to extract parameters from simulation data that can be directly used for fracture mechanics evaluation, thereby quantitatively predicting the tendency for crack initiation and propagation.

[0039] Specifically, the first step is to identify regions in the metal components where the principal stress exceeds the material's yield strength, and record these regions as stress concentration areas. Principal stresses refer to the main stress components of a metal material caused by external loads or thermal stress during the cutting process. When the principal stresses exceed the material's yield strength, the material enters the plastic deformation zone, undergoing permanent deformation, and may even lead to crack initiation. For example, all solid elements are traversed, and the calculated principal stresses at each integration point—the first, second, and third principal stresses—are read. These are compared with the material yield strength, corresponding to the current element temperature and phase composition, retrieved in real-time from the material property database. The system sets a criterion: when the first principal stress of an element, i.e., the maximum tensile stress, exceeds the material yield strength of that element in its current state, the element is marked as a high-risk element. At each future time step, all marked high-risk elements are spatially clustered to form one or more continuous volumetric regions, which are recorded as stress concentration areas for the current and future times. This identification process continues throughout the entire future time window, allowing tracking of the evolution of stress concentration areas as the cutting process progresses, including dynamic changes in their location, size, and stress level.

[0040] Subsequently, for each identified stress concentration region, refined calculations of fracture mechanics parameters are performed. Since cracks can initiate and propagate in different directions, this step considers multiple pre-defined potential propagation directions. These pre-defined directions are typically determined based on the stress state of the region and the crystallographic properties or macroscopic defect orientation of the material; for example, they can be pre-defined as directions perpendicular to the maximum principal stress, or directions forming a specific angle with it. For each high-risk element on the boundary of the stress concentration region, or for the virtual crack front defined for that region, based on detailed stress and displacement field information near that location provided by advanced simulation data, numerical methods are used to calculate the stress intensity factor for three basic modes under different pre-defined propagation directions. For example, the Type I stress intensity factor corresponding to the opening crack propagation mode is calculated using the interaction integral method or displacement extrapolation method based on the normal stress distribution along the virtual crack propagation direction, to characterize the ability of tensile stress perpendicular to the crack surface to drive crack opening; the Type II stress intensity factor corresponding to the sliding crack propagation mode is calculated using the interaction integral method based on the in-plane shear stress distribution along the virtual crack propagation direction, to characterize the ability of shear stress parallel to the crack surface and perpendicular to the crack tip to drive in-plane sliding; and the Type III stress intensity factor corresponding to the tearing crack propagation mode is calculated using the interaction integral method based on the out-plane shear stress distribution along the virtual crack propagation direction, to characterize the ability of shear stress parallel to the crack surface and parallel to the crack tip to drive out-plane tearing.

[0041] Next, the overall risk of crack initiation and propagation under combined stress conditions is further assessed. Since actual defects are often in combined stress states, the stress intensity factors for the three modes mentioned above need to be combined. For example, an equivalent stress intensity factor can be calculated based on the maximum circumferential stress criterion, the maximum energy release rate criterion, or empirical formulas. Simultaneously, the dynamic fracture toughness of the material, matching the average strain rate, temperature, and phase composition of the current stress concentration region, is obtained from a material dynamic property database. This dynamic fracture toughness considers the changes in the material's fracture resistance under high loading rates.

[0042] Finally, the ratio of the composite stress intensity factor to the dynamic fracture toughness of the material in each preset propagation direction is calculated. This ratio is a dimensionless risk coefficient. When the ratio is less than 1, it indicates that the predicted driving force is less than the material's fracture resistance, and the risk is low. When the ratio is close to or equal to 1, it indicates that a critical state has been reached. When the ratio is greater than 1, it indicates that the predicted driving force has exceeded the material's resistance, and cracks may initiate and propagate unstablely. The system can take the maximum value of this ratio in all stress concentration regions and all preset propagation directions as a global potential defect initiation risk indicator. Alternatively, for each critical region, the curve of the maximum ratio changing over time can be recorded to assess the development trend of risk as the cutting process progresses, providing a clear basis for judging the safety of the processing.

[0043] P30: Based on the deformation trend of the component and the potential risk of defect emergence, perform multi-objective optimization to determine the optimal composite intervention strategy for the next time step.

[0044] Furthermore, step P30 in this embodiment of the application also includes: P31: Calculate the contour error evaluation value based on the deformation trend of the component, calculate the crack risk evaluation value based on the potential defect initiation risk, and obtain the current cutting efficiency evaluation value; P32: Construct a multi-objective optimization function with the optimization objectives of minimizing the contour error evaluation value, minimizing the crack risk evaluation value, and maximizing the current cutting efficiency evaluation value; P33: Using energy source parameters, motion axis path parameters, and auxiliary medium parameters as decision variables, solve the multi-objective optimization function in real time under the constraints of the processing technology, obtain the combination of decision variables that makes the multi-objective optimization function value optimal, and generate the optimal composite intervention strategy for the next time step.

[0045] Optionally, based on two core predictive indicators—component deformation trends and the risk of potential defect initiation—a multi-objective optimization decision-making process is used to generate an optimal composite intervention strategy capable of coordinating the control of multiple actuators to dynamically respond to changes in the mechanical state during processing. This strategy is not merely an adjustment of a single parameter, but rather a collaborative planning of multiple control dimensions, including energy sources, motion axes, and auxiliary media, to achieve comprehensive optimization of accuracy, quality, and efficiency.

[0046] Specifically, the predicted physical quantities are first transformed into calculable optimization target evaluation values. Based on the component deformation trend, i.e., the curve of the root mean square value of the full-field displacement normal deviation changing over time within the future time window, it is quantified into a scalar profile error evaluation value through an evaluation function. This evaluation function can be the instantaneous value of the root mean square value at a key future time point (such as the end of the prediction window), or its integral value or maximum value over the entire prediction time window; the larger the value, the more severe the predicted shape distortion. Simultaneously, based on the potential defect initiation risk quantification index, i.e., the ratio of the composite stress intensity factor to the dynamic fracture toughness or its evolution curve, a crack risk evaluation value is calculated through another evaluation function. This function can also take the maximum value, the final value of the window, or the integral value of this ratio; the larger the value, the higher the risk of crack initiation and propagation. Furthermore, it is necessary to obtain the current cutting efficiency evaluation value. The cutting efficiency evaluation value reflects the relationship between the time, energy, and other resources consumed in the current cutting process and the achieved cutting effect; higher cutting efficiency indicates more optimized resource utilization.

[0047] Subsequently, a multi-objective optimization function is constructed with the optimization objectives of minimizing the contour error evaluation value, minimizing the crack risk evaluation value, and maximizing the current cutting efficiency evaluation value. The purpose of this function is to balance the three objectives in the cutting process: reducing workpiece geometric errors, reducing crack risk, and improving cutting efficiency. To achieve this objective, the multi-objective optimization function will combine these three objective values ​​and be constructed using a weighted summation method or the Pareto optimality method. For example, higher weights can be assigned to contour error and crack risk to ensure the priority of processing quality, while an appropriate weight can be assigned to cutting efficiency to ensure the efficiency of the processing process. The final multi-objective optimization function will consider these objectives simultaneously and find a balance point that satisfies all requirements.

[0048] Furthermore, adjustable processing parameters are used as decision variables, and the aforementioned multi-objective optimization function is solved in real time under the physical constraints of the process and equipment. Specific decision variables include: energy source parameters, such as laser power, pulse frequency, pulse width, and defocusing amount in laser cutting; motion axis path parameters, mainly referring to the instantaneous velocity vector of the cutting head in the next time step, and optional small lateral or longitudinal offsets made locally to the preset path; auxiliary medium parameters, such as the pressure and flow rate of auxiliary gases (e.g., oxygen, nitrogen), or the on / off state, flow rate, or temperature of auxiliary cooling jets used for active thermal management. Process constraints include physical upper and lower limits for each decision variable (e.g., laser power must not exceed the maximum value of the equipment), maximum acceleration and jerk limits for each motion axis, and geometric constraints such as avoiding collisions with the fixture.

[0049] The solution process is performed online within each control cycle. The system calls upon an optimization algorithm library, such as a constrained non-dominated sorting genetic algorithm or a particle swarm optimization algorithm, to rapidly find the optimal solution within the search space comprised of decision variables, guided by the current objective function F. The optimization algorithm evaluates a large number of candidate decision variable combinations. For each set of candidate parameters, rapid forward simulation is performed based on a real-time updated digital twin to predict the profile error and crack risk within a future time window under that set of parameters, and the objective function F is calculated in conjunction with its inherent efficiency value. Ultimately, the algorithm finds the set of decision variables that optimizes the objective function F. This set defines the specific parameter settings that the energy source, motion axis, and auxiliary medium should coordinately adjust to in the next control time step, thus constituting the optimal composite intervention strategy. This strategy is then deployed to the equipment execution layer to achieve closed-loop adaptive and precise control of the processing process.

[0050] Furthermore, in generating the optimal composite intervention strategy for the next time step, step P33 of this embodiment also includes: P33-1: When the potential defect initiation risk exceeds a first risk threshold, a first intervention strategy is generated with the core optimization objective of suppressing crack initiation. The first intervention strategy includes switching the continuous energy source to an ultra-short pulse mode and controlling the cutting head to execute a micro-oscillation path that locally widens the cutting kerf. P33-2: When the component deformation trend exceeds a first deformation threshold and the potential defect initiation risk does not exceed the first risk threshold, a second intervention strategy is generated with the core optimization objective of minimizing the final contour error. The second intervention strategy includes introducing a reverse pre-compensation offset in the subsequent cutting path based on the component deformation trend. P33-3: When the component deformation trend does not exceed the first deformation threshold and the potential defect initiation risk does not exceed the first risk threshold, a third intervention strategy is generated with the core optimization objective of maximizing cutting efficiency. The third intervention strategy includes dynamically increasing the cutting feed speed under preset processing quality constraints.

[0051] It should be understood that, based on the solution results of the multi-objective optimization function, in order to adapt to different emergency risk levels and process target priorities, this application further defines a more targeted strategy generation logic under specific threshold triggering conditions. Under the premise of ensuring that the core risks are controllable, it optimizes the real-time performance and execution efficiency of decision-making and generates the optimal composite intervention strategy for the next time step.

[0052] Specifically, a first risk threshold related to the material's fracture toughness and safety factor is set. When the quantitative index of potential defect initiation risk calculated based on advanced simulation data, such as the ratio of composite stress intensity factor to dynamic fracture toughness, exceeds this first risk threshold, the system determines that there is an imminent risk of crack initiation or unstable propagation. In this case, the system switches its decision logic to a mode with crack initiation suppression as the core optimization objective and generates a first intervention strategy. This strategy is formulated based on fracture mechanics and process knowledge base, and its core is to immediately reduce the thermal concentration effect of energy input and change the local stress state. Specific measures include: First, switching the continuous energy source to an ultra-short pulse mode. This operation transforms continuous high energy input into high-frequency, short-pulse pulsed energy input. While the total average power may decrease, extremely high peak power is used to achieve material vaporization and removal. However, the duration of each pulse is extremely short, thereby significantly reducing the heat conducted into the workpiece, reducing the heat-affected zone and thermal stress. At the same time, the interval between pulses provides cooling time for the material, which is beneficial for reducing the average temperature. Secondly, the cutting head is synchronously controlled to execute a micro-oscillation path that locally widens the kerf. That is, while maintaining the overall feed direction, the cutting head is controlled to oscillate periodically in a plane perpendicular to the feed direction with high frequency and small amplitude (e.g., an amplitude 1 to 2 times the spot diameter), such as a circular, figure-eight, or sinusoidal trajectory. This micro-oscillation path physically widens the kerf, directly releasing local constraints near the cut, providing additional space for stress redistribution, and reducing stress concentration. Simultaneously, the wider kerf reduces the risk of interference between the cutting head and the already cut sidewalls. This first intervention strategy is risk-averse, prioritizing the reliability of the processing and preventing catastrophic failure.

[0053] Simultaneously, a first deformation threshold is set based on the final part accuracy requirements. When the predicted component deformation trend, such as the root mean square value of the contour error at the end of a future time window, exceeds this first deformation threshold, but the quantification index of potential defect initiation risk does not exceed the first risk threshold, the current main problem is determined to be deformation exceeding the limit, but there is no immediate risk of cracking. At this time, the decision logic takes minimizing the final contour error as the core optimization objective and generates a second intervention strategy. The core idea of ​​this strategy is proactive compensation. The specific measures are: introducing a reverse pre-compensation offset in the subsequent cutting path based on the component deformation trend. The system calculates the drift error of the cutting path in physical space caused by the deformation of the machined and unmachined parts of the workpiece at future moments based on the deformation field predicted by the digital twin. Then, when generating the CNC command for the subsequent cutting path, the theoretical geometric path is corrected in real time and dynamically. The magnitude of the correction is equal to the predicted deformation displacement, but in the opposite direction, so that theoretically the actual cutting trajectory can fall exactly on the ideal part contour after deformation drift. This path correction is a geometric feedforward compensation that does not change the root causes of heat input and stress generation, but offsets their ultimate impact on shape accuracy through precise trajectory adjustments. When implementing this strategy, energy source parameters and auxiliary medium parameters can be maintained at their usual optimized settings or fine-tuned to assist in deformation control.

[0054] Furthermore, when the predicted component deformation trend does not exceed the first deformation threshold, and the potential defect initiation risk also does not exceed the first risk threshold, the current processing state is determined to be stable and within the safety and accuracy margins. At this point, the decision logic shifts to maximizing cutting efficiency as the core optimization objective, generating a third intervention strategy. This strategy aims to maximize material removal rate and shorten processing time while ensuring that processing quality meets preset constraints. Specifically, under preset processing quality constraints, the cutting feed rate is dynamically increased. Preset processing quality constraints typically include maximum surface roughness and kerf width tolerance. At each decision, the system attempts to propose a higher feed rate candidate value based on the current speed, either by a fixed step size (e.g., increasing by 5 mm per second) or within the optimization algorithm framework. Then, the system uses a digital twin to quickly evaluate whether the predicted contour error and crack risk at the increased speed are still below the first deformation threshold and the first risk threshold, respectively. If all constraints are met, the higher feed rate is adopted as part of the optimal strategy. Simultaneously, parameters such as energy source power may be adjusted upwards to maintain sufficient energy density to ensure cutting continuity. This strategy is a dynamic optimization process that continuously explores the upper limit of efficiency that can be achieved under the current working conditions, thereby achieving adaptive and efficient processing.

[0055] Through the threshold-based condition judgment and strategy selection mechanism described above, the system can intelligently and agilely switch and weigh the three core objectives of risk control, accuracy assurance, and efficiency improvement, thereby generating an optimal composite intervention strategy that is both safe and reliable as well as cost-effective.

[0056] P40: Control the cutting equipment to execute the optimal composite intervention strategy, and after execution, return to the original location to collect sensing data for closed-loop adaptive control.

[0057] Furthermore, step P40 in this embodiment of the application also includes: P41: Decompose the optimal composite intervention strategy into instructions to generate a synchronized time-energy-trajectory coordinated control sequence; P42: Send the energy control instructions from the time-energy-trajectory coordinated control sequence to the energy source controller, the trajectory control instructions to the multi-axis motion controller, and the media control instructions to the auxiliary media controller; P43: Triggering the synchronized time-energy-trajectory coordinated control sequence, control the energy source, the multi-axis motion controller, and the auxiliary media controller to perform coordinated actions; P44: During the execution of the coordinated actions, synchronously record the actual coordinated action data and feed the coordinated action data back to the process-level digital twin to correct the actuator response model in the coupled field model.

[0058] Specifically, controlling the cutting equipment to execute the optimal composite intervention strategy and then performing closed-loop adaptive control after execution are key steps to ensure the cutting process is precise, efficient, and meets expectations. Through the execution of the optimal composite intervention strategy and real-time feedback, various parameters in the cutting process can be dynamically adjusted and optimized.

[0059] First, the optimal composite intervention strategy is decomposed into commands to generate a synchronized time-energy-trajectory coordinated control sequence. The optimal composite intervention strategy typically includes multiple control parameters, such as the control of the energy source, the motion trajectory of the cutting head, and the adjustment of the auxiliary medium. To ensure that these control parameters work collaboratively, the system first decomposes the optimal composite intervention strategy into different control commands and synchronizes them according to time sequence, energy demand, and motion trajectory requirements. This control sequence precisely specifies how much energy the energy source should output, what path the motion control system should take, and how the auxiliary medium should be adjusted within each time step. Through this command decomposition, it ensures that each control module can execute its task at the correct time and under the correct conditions, thereby achieving global optimization.

[0060] Next, the control commands in the generated time-energy-trajectory coordinated control sequence are sent to each controller. Energy control commands are sent to the energy source controller, which adjusts the output of energy sources such as laser power and cutting airflow to achieve precise cutting energy input. Trajectory control commands are sent to the multi-axis motion controller to ensure the cutting head moves along a predetermined path and speed, guaranteeing cutting accuracy. Medium control commands are sent to the auxiliary medium controller to adjust parameters such as the flow rate and pressure of gases and liquids during the cutting process, ensuring effective cooling and heat removal. By sending these control commands to each controller, the system can coordinately control all aspects of the cutting process.

[0061] Finally, based on the triggering of the synchronized time-energy-trajectory coordinated control sequence, the energy source, multi-axis motion controller, and auxiliary media controller are controlled to perform coordinated actions. During the actual cutting process, the system precisely controls the energy source, motion control system, and auxiliary media controller to work synchronously according to the trigger signal of the time-energy-trajectory coordinated control sequence. For example, when the energy source outputs a specific power, the multi-axis motion controller starts and moves along a preset path, while the auxiliary media controller adjusts the flow rate according to the settings to ensure that the temperature and environmental conditions of the cutting head are always at their optimal levels. This coordinated action ensures the smooth progress of the cutting process and avoids cutting errors or instability caused by incoordination between control modules.

[0062] Furthermore, during the execution of coordinated actions, the system synchronously records the actual data of the coordinated actions and feeds this data back to the process-level digital twin to correct the actuator response model in the coupled field model. By acquiring real-time data on the actual coordinated actions, such as actual energy output, actual cutting path, and auxiliary medium flow rate, the system can compare the differences between the predetermined control target and the actual execution result and make corrections. This data will be fed back to the process-level digital twin to update the actuator response model in the coupled field model. The actuator response model simulates the relationship between the controller and the actuator, including the dynamic response between the control signal and the actual execution action. By correcting this model, the system can more accurately predict future cutting processes and adjust the control strategy in real time to ensure the accuracy and stability of the cutting process.

[0063] Through these steps, the system can achieve comprehensive control over the cutting process, ensuring not only the coordinated operation of the energy source, motion control system, and auxiliary medium, but also continuously optimizing the control strategy through a closed-loop feedback mechanism, thereby improving the quality and efficiency of the cutting process.

[0064] In summary, the embodiments of this application have at least the following technical effects: This application achieves precise control over component deformation trends and potential defect initiation risks by real-time monitoring of vibration, acoustic emission, thermal effects, and microstructure of metal components during the cutting process. This allows for dynamic adjustment of the cutting strategy, ensuring no excessive deformation, cracks, or dimensional deviations occur during cutting, significantly improving processing accuracy. Real-time sensing and prediction of residual stress generated during cutting effectively prevents crack formation and material failure caused by thermal stress changes and stress concentration, enhancing the stability of the cutting process. Based on a multi-objective optimization algorithm, key parameters such as energy source, cutting path, and auxiliary medium can be adjusted in real-time to achieve an efficient cutting process and minimize potential defect risks and processing errors. By real-time acquisition of in-situ sensing data and feedback to a digital twin, closed-loop control can be implemented, continuously adjusting and optimizing the cutting process to improve cutting efficiency and workpiece quality.

[0065] It achieves the technical effect of closed-loop adaptive control based on multi-domain perception and digital twin, which can predict and actively compensate for deformation and defects caused by stress release in real time, thereby improving the cutting accuracy and stability of metal parts under complex stress conditions.

[0066] Example 2, based on the same inventive concept as the cutting control method for metal parts processing in the foregoing examples, such as... Figure 2 As shown, this application provides a cutting control system for processing metal parts. The system and method embodiments in this application are based on the same inventive concept. The system includes: The data synchronization acquisition module 11 is used to synchronously and in real time acquire in-situ sensing data of at least two physical domains during the movement of the cutting point. The physical domains include the vibration domain, acoustic emission domain, thermal domain, and micromorphological domain.

[0067] The cutting risk prediction module 12 is used to drive and update the process-level digital twin of the metal parts in real time based on the in-situ sensing data, solve the coupled field model of the process-level digital twin, and predict the component deformation trend and potential defect initiation risk of the current cutting action in the future time window. The process-level digital twin is coupled with the structural mechanical field, the heat conduction field and the material phase transition field.

[0068] The multi-objective optimization solution module 13 is used to perform multi-objective optimization based on the deformation trend of the component and the potential defect initiation risk, and to decide and generate the optimal composite intervention strategy for the next time step.

[0069] The cutting execution feedback module 14 is used to control the cutting equipment to execute the optimal composite intervention strategy, and after execution, return to the original location to collect sensing data for closed-loop adaptive control.

[0070] Furthermore, the data synchronization acquisition module 11 is also used to perform the following steps: The microscopic vibration signal of the processed area behind the cutting point is acquired in real time by a laser Doppler vibrometer to obtain in-situ sensing data in the vibration domain; the elastic wave signal excited by the microscopic fracture of the material during the cutting process is acquired in real time by an acoustic emission sensor array to obtain in-situ sensing data in the acoustic emission domain; the dynamic temperature field signal including the cutting seam and the heat-affected zone is acquired in real time by a high-speed infrared thermal imager to obtain in-situ sensing data in the thermal domain; and the microscopic geometric morphology signal of the cutting seam sidewall is acquired in real time by an optical coherence tomography module to obtain in-situ sensing data in the micromorphology domain.

[0071] Furthermore, the cutting risk prediction module 12 is also used to perform the following steps: Based on the geometric, material, and constraint information of the metal components, a basic digital twin containing an initial residual stress field is constructed. Within this basic digital twin, a coupled field model is configured, integrating a heat source model, a heat transfer model, a microstructure evolution model, and a stress-strain calculation model. The heat source model maps the control parameters of the cutting energy source to heat flux density distribution parameters applied to the workpiece surface. Based on the in-situ sensing data, the basic digital twin is updated in real time, including: Based on the in-situ thermal sensing data and the current energy source control parameters, the heat flux density distribution parameters of the heat source model are inverted in real time to calibrate the energy input. Based on the calibrated energy input, combined with the in-situ thermal sensing data, the thermophysical parameters of the heat transfer model are inverted and calibrated in real time to update the heat conduction field. Through the in-situ acoustic emission domain sensing data and the in-situ microstructure domain sensing data, material damage characteristics are identified, and the tissue evolution model is corrected in real time to update the material phase transition field. Based on the in-situ vibration domain sensing data and the updated heat conduction field and material phase transition field, the stress-strain calculation model is corrected in real time to update the structural mechanical field.

[0072] Furthermore, the cutting risk prediction module 12 is also used to perform the following steps: Based on the coupled field model after real-time inversion calibration and real-time feedback correction using in-situ sensing data, advanced numerical simulation is performed, and the advanced simulation data is recorded. According to the advanced simulation data, the full-field displacement data of the metal components within a future time window is extracted, and the root mean square value of the normal deviation of the full-field displacement data relative to the ideal geometric profile is calculated to obtain the deformation trend of the component. According to the advanced simulation data, the stress intensity factor of the stress concentration region in the metal components within a future time window is extracted, and the safety margin of the stress intensity factor relative to the material fracture toughness is calculated. The reciprocal of the safety margin is used as the potential defect initiation risk.

[0073] Furthermore, the cutting risk prediction module 12 is also used to perform the following steps: During the advanced simulation process, regions in the metal components where the principal stress exceeds the material's yield strength are identified and recorded as stress concentration regions. For each stress concentration region, based on the advanced simulation data, Type I, Type II, and Type III stress intensity factors are calculated in different preset propagation directions. The Type I, Type II, and Type III stress intensity factors are combined to obtain a composite stress intensity factor for the stress concentration region, and the ratio of the composite stress intensity factor to the material's dynamic fracture toughness is calculated as a quantitative indicator of the potential defect initiation risk.

[0074] Furthermore, the multi-objective optimization solution module 13 is also used to perform the following steps: Based on the deformation trend of the component, the contour error evaluation value is calculated, and based on the potential defect initiation risk, the crack risk evaluation value is calculated, and the current cutting efficiency evaluation value is obtained. A multi-objective optimization function is constructed with the optimization objectives of minimizing the contour error evaluation value, minimizing the crack risk evaluation value, and maximizing the current cutting efficiency evaluation value. Using energy source parameters, motion axis path parameters, and auxiliary medium parameters as decision variables, the multi-objective optimization function is solved in real time under processing constraints to obtain the optimal combination of decision variables that optimizes the multi-objective optimization function value, and the optimal composite intervention strategy for the next time step is generated.

[0075] Furthermore, the multi-objective optimization solution module 13 is also used to perform the following steps: When the potential defect initiation risk exceeds a first risk threshold, a first intervention strategy is generated with the core optimization objective of suppressing crack initiation. The first intervention strategy includes switching the continuous energy source to an ultra-short pulse mode and controlling the cutting head to execute a micro-oscillation path that locally widens the cutting kerf. When the component deformation trend exceeds a first deformation threshold and the potential defect initiation risk does not exceed the first risk threshold, a second intervention strategy is generated with the core optimization objective of minimizing the final contour error. The second intervention strategy includes introducing a reverse pre-compensation offset in the subsequent cutting path based on the component deformation trend. When the component deformation trend does not exceed the first deformation threshold and the potential defect initiation risk does not exceed the first risk threshold, a third intervention strategy is generated with the core optimization objective of maximizing cutting efficiency. The third intervention strategy includes dynamically increasing the cutting feed speed under preset processing quality constraints.

[0076] Furthermore, the cutting execution feedback module 14 is also used to perform the following steps: The optimal composite intervention strategy is decomposed into instructions to generate a synchronized time-energy-trajectory coordinated control sequence. Energy control instructions from this sequence are sent to the energy source controller, trajectory control instructions to the multi-axis motion controller, and media control instructions to the auxiliary media controller. Triggered by the synchronized time-energy-trajectory coordinated control sequence, the energy source, multi-axis motion controller, and auxiliary media controller are controlled to perform coordinated actions. During the execution of these coordinated actions, the actual coordinated action data is recorded synchronously and fed back to the process-level digital twin to correct the actuator response model in the coupled field model.

[0077] Example 3: Exemplary electronic device.

[0078] The following is for reference. Figure 3 The present application describes the electronic device according to its embodiments.

[0079] Based on the same inventive concept as the cutting control method for metal parts processing in the foregoing embodiments, this application also provides a cutting control system for metal parts processing, including: a processor coupled to a memory for storing a program, wherein when the program is executed by the processor, the system performs the steps of the method described in Embodiment 1.

[0080] The electronic device 300 includes a processor 302, a communication interface 303, and a memory 301. Optionally, the electronic device 300 may also include a bus architecture 304. The communication interface 303, processor 302, and memory 301 can be interconnected via the bus architecture 304; the bus architecture 304 can be a peripheral component interconnect (PCI) bus or an extended industry standard architecture (EISA) bus, etc. The bus architecture 304 can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 3 The bus is represented by a single thick line, but this does not mean that there is only one bus or one type of bus.

[0081] Processor 302 may be a CPU, microprocessor, ASIC, or one or more integrated circuits used to control the execution of programs according to the present application.

[0082] Communication interface 303 uses any transceiver-like device for communicating with other devices or communication networks, such as Ethernet, radio access network (RAN), wireless local area network (WLAN), wired access network, etc.

[0083] Memory 301 can be ROM or other types of static storage devices capable of storing static information and instructions, RAM or other types of dynamic storage devices capable of storing information and instructions, or electrically erasable programmable read-only memory (EEPROM), compact disc read-only memory (CDROM) or other optical disc storage, optical disc storage (including compressed optical discs, laser discs, optical discs, digital universal optical discs, Blu-ray discs, etc.), magnetic disk storage media or other magnetic storage devices, or any other medium capable of carrying or storing desired program code in the form of instructions or data structures and accessible by a computer, but not limited thereto. Memory can exist independently and be connected to the processor via bus architecture 304. Memory can also be integrated with the processor.

[0084] The memory 301 stores computer execution instructions for implementing the scheme of this application, and the processor 302 controls the execution. The processor 302 executes the computer execution instructions stored in the memory 301, thereby implementing the cutting control method for metal parts processing provided in the above embodiments of this application.

[0085] It should be noted that the order of the embodiments described above is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. Furthermore, the above description focuses on specific embodiments of this specification. Additionally, the processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired results. In some implementations, multitasking and parallel processing are possible or may be advantageous.

[0086] The above description is only a preferred embodiment of this application and is not intended to limit this application. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.

[0087] This specification and accompanying drawings are merely illustrative examples of this application and are intended to cover any and all modifications, variations, combinations, or equivalents within the scope of this application. Clearly, those skilled in the art can make various alterations and modifications to this application without departing from its scope. Therefore, if such modifications and variations fall within the scope of this application and its equivalents, this application intends to include such modifications and variations.

Claims

1. A cutting control method for machining metal parts, characterized in that, The method includes: During the movement of the cutting point, in-situ sensing data from at least two physical domains are collected simultaneously in real time. These physical domains include the vibration domain, acoustic emission domain, thermal domain, and micromorphological domain. Based on the in-situ sensing data, the process-level digital twin of the metal parts is driven and updated in real time. The coupled field model of the process-level digital twin is solved to predict the deformation trend of the component and the risk of potential defect initiation in the future time window of the current cutting action. The process-level digital twin is coupled with the structural mechanical field, the heat conduction field and the material phase transition field. Based on the deformation trend of the component and the risk of potential defect emergence, a multi-objective optimization solution is performed to generate the optimal composite intervention strategy for the next time step. The cutting equipment is controlled to execute the optimal composite intervention strategy, and after execution, it returns to the original location to collect sensing data for closed-loop adaptive control.

2. The cutting control method for metal parts processing as described in claim 1, characterized in that, During the movement of the cutting point, a synchronous acquisition timestamp is set, and synchronous data acquisition is performed based on the synchronous acquisition timestamp, including: The microscopic vibration signal of the processed area behind the cutting point is collected in real time by a laser Doppler vibration meter to obtain in-situ sensing data of the vibration domain. The acoustic emission sensor array is used to collect elastic wave signals excited by the microscopic fracture of the material during the cutting process in real time, so as to obtain in-situ sensing data in the acoustic emission domain. The dynamic temperature field signal, including the cutting seam and heat-affected zone, is acquired in real time by a high-speed infrared thermal imager to obtain in-situ thermal sensing data. The microscopic geometric morphology signal of the cut seam sidewall is acquired in real time by an optical coherence tomography module to obtain in-situ sensing data of the microscopic morphology domain.

3. The cutting control method for metal parts processing as described in claim 2, characterized in that, Based on the in-situ sensing data, the process-level digital twin of the metal parts is driven and updated in real time, including: Based on the geometry, material and constraint information of metal parts, a basic digital twin containing the initial residual stress field is constructed; In the basic digital twin, a coupled field model is configured that couples a heat source model, a heat transfer model, a microstructure evolution model, and a stress-strain calculation model. The heat source model is used to map the control parameters of the cutting energy source to the heat flux density distribution parameters applied to the workpiece surface. Based on the in-situ sensing data, the basic digital twin is updated in real time, including: Based on the in-situ sensing data of the thermal domain and the current energy source control parameters, the heat flux density distribution parameters of the thermal source model are inverted in real time to calibrate the energy input; Based on the calibrated energy input, combined with the in-situ sensing data of the thermal domain, the thermophysical parameters of the heat transfer model are inverted and calibrated in real time to update the heat conduction field. By using the in-situ sensing data in the acoustic emission domain and the in-situ sensing data in the micromorphology domain, material damage characteristics are identified, and the tissue evolution model is corrected in real time to update the material phase transition field. Based on the in-situ sensing data of the vibration domain and the updated heat conduction field and material phase transition field, the stress-strain calculation model is corrected in real time to update the structural mechanical field.

4. The cutting control method for metal parts processing as described in claim 3, characterized in that, Solving the coupled field model of the process-level digital twin to predict the component deformation trend and potential defect initiation risk within a future time window after the current cutting action includes: Based on the coupled field model after real-time inversion calibration and real-time feedback correction using in-situ sensing data, advanced numerical simulation is performed, and the advanced simulation data is recorded. Based on the advanced simulation data, the full-field displacement data of the metal parts within the future time window is extracted, and the root mean square value of the normal deviation of the full-field displacement data relative to the ideal geometric profile is calculated to obtain the deformation trend of the component. Based on the advanced simulation data, the stress intensity factor of the stress concentration region in the metal component within the future time window is extracted, and the safety margin of the stress intensity factor relative to the material fracture toughness is calculated. The reciprocal of the safety margin is taken as the potential defect initiation risk.

5. The cutting control method for metal parts processing as described in claim 4, characterized in that, Based on the advanced simulation data, the stress intensity factor of the stress concentration region in the metal component within the future time window is extracted to obtain the potential defect initiation risk, which also includes: During the advanced simulation process, regions in the metal parts where the principal stress exceeds the material yield strength are identified and recorded as stress concentration regions. For each stress concentration region, based on the aforementioned advanced simulation data, calculate its Type I stress intensity factor, Type II stress intensity factor, and Type III stress intensity factor in different preset propagation directions; The Type I, Type II, and Type III stress intensity factors are combined to obtain a composite stress intensity factor for the stress concentration region. The ratio of the composite stress intensity factor to the dynamic fracture toughness of the material is calculated as a quantitative indicator of the potential defect initiation risk.

6. The cutting control method for metal parts processing as described in claim 1, characterized in that, Based on the deformation trend of the component and the potential risk of defect initiation, a multi-objective optimization solution is performed to determine the optimal composite intervention strategy for the next time step, including: The contour error evaluation value is calculated based on the deformation trend of the component, the crack risk evaluation value is calculated based on the potential defect initiation risk, and the current cutting efficiency evaluation value is obtained. A multi-objective optimization function is constructed with the optimization objectives of minimizing the contour error evaluation value, minimizing the crack risk evaluation value, and maximizing the current cutting efficiency evaluation value. Using energy source parameters, motion axis path parameters, and auxiliary medium parameters as decision variables, the multi-objective optimization function is solved in real time under the constraints of the processing technology. The combination of decision variables that makes the value of the multi-objective optimization function optimal is obtained, and the optimal composite intervention strategy for the next time step is generated.

7. The cutting control method for metal parts processing as described in claim 6, characterized in that, Generate the optimal composite intervention strategy for the next time step, including: When the potential defect initiation risk exceeds the first risk threshold, a first intervention strategy is generated with the core optimization goal of suppressing crack initiation. The first intervention strategy includes switching the continuous energy source to an ultra-short pulse mode and controlling the cutting head to execute a micro-oscillation path that locally widens the cutting kerf. When the deformation trend of the component exceeds the first deformation threshold and the risk of potential defect occurrence does not exceed the first risk threshold, a second intervention strategy is generated with minimizing the final contour error as the core optimization objective. The second intervention strategy includes introducing a reverse pre-compensation offset in the subsequent cutting path based on the deformation trend of the component. When the deformation trend of the component does not exceed the first deformation threshold and the risk of potential defect occurrence does not exceed the first risk threshold, a third intervention strategy is generated with maximizing cutting efficiency as the core optimization objective. The third intervention strategy includes dynamically increasing the cutting feed speed under preset processing quality constraints.

8. The cutting control method for metal parts processing as described in claim 1, characterized in that, Controlling the cutting equipment to execute the optimal composite intervention strategy includes: The optimal composite intervention strategy is decomposed into instructions to generate a synchronized time-energy-trajectory coordinated control sequence; The energy control command in the time-energy-trajectory coordinated control sequence is sent to the energy source controller, the trajectory control command is sent to the multi-axis motion controller, and the medium control command is sent to the auxiliary medium controller; Under the triggering of the synchronized time-energy-trajectory coordinated control sequence, the energy source, the multi-axis motion controller, and the auxiliary medium controller are controlled to perform coordinated actions; During the execution of the collaborative action, the actual data of the collaborative action is recorded synchronously, and the collaborative action data is fed back to the process-level digital twin to correct the actuator response model in the coupled field model.

9. A cutting control system for processing metal parts, characterized in that, The system includes: The data synchronization acquisition module is used to synchronously and in real time acquire in-situ sensing data of at least two physical domains during the movement of the cutting point. The physical domains include the vibration domain, acoustic emission domain, thermal domain, and micromorphological domain. The cutting risk prediction module is used to drive and update the process-level digital twin of metal parts in real time based on the in-situ sensing data, solve the coupled field model of the process-level digital twin, and predict the component deformation trend and potential defect initiation risk of the current cutting action in the future time window. The process-level digital twin is coupled with the structural mechanical field, the heat conduction field and the material phase transition field. The multi-objective optimization solution module is used to perform multi-objective optimization based on the deformation trend of the component and the potential defect initiation risk, and to generate the optimal composite intervention strategy for the next time step. The cutting execution feedback module is used to control the cutting equipment to execute the optimal composite intervention strategy, and after execution, return to the original location to collect sensing data for closed-loop adaptive control.

10. An electronic device, characterized in that, include: A processor coupled to a memory for storing a program that, when executed by the processor, causes the system to perform the steps of the method as claimed in any one of claims 1 to 8.

Images