Laser spot regulation and control system with self-adaptive feedback mechanism and regulation and control method thereof

Abstract

The invention is suitable for the technical field of laser spot regulation and control, and provides a laser spot regulation and control system with a self-adaptive feedback mechanism and a regulation and control method thereof, and the system comprises a workpiece detection module, a model generation module, a learning analysis module, a working condition collection module, an optimization adjustment module and a hardware execution module. The workpiece detection module is used for collecting workpiece space data and workpiece material components; the model generation module is used for generating a digital twinborn body according to the workpiece space data and the workpiece material components; the learning analysis module is used for analyzing the digital twins and generating or optimizing process parameters, working paths and to-be-processed data in cooperation with the self-learning processing library; the working condition collecting module is used for collecting actual machining data and external environment data in the workpiece machining process. According to the device, the problem of insufficient laser spot regulation and control precision is solved, and the effects of reducing deviation and improving the machining precision are achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of laser spot control, and more particularly, to a laser spot control system with an adaptive feedback mechanism and a control method thereof. Background Art

[0002] Since its inception, laser technology has been widely used in scientific research, industry, medicine, communications, military, and daily life. Today, precise control of laser spot parameters has become a core element of laser technology.

[0003] At present, there are many factors that are known to affect the accuracy of the laser spot. First, the environmental sensitivity problem. When the working environment temperature fluctuates by more than ±5℃, the laser focus may drift by 0.1mm-0.3mm. In addition, equipment operating in common industrial environments will generate equipment vibration, and its vibration frequency can reach 5-200Hz. These factors will significantly change the effective focal depth, resulting in a large deviation rate of the working depth; second, the workpiece material problem. Some workpieces need to select two or even more different materials for processing. The surface reflectivity of the workpieces varies greatly, which will also affect the focal position of the laser; third, the dynamic response problem. The existing laser spot control device or system that can realize automatic correction can reduce some errors, but it cannot accurately compensate for some nonlinear or uncertain errors, which still affects the accuracy of the laser spot.

[0004] Therefore, a laser spot control system with an adaptive feedback mechanism and a control method thereof are proposed to solve the above problems. Summary of the Invention

[0005] In view of the shortcomings of the prior art, the object of the present invention is to provide a laser spot control system and a control method thereof with an adaptive feedback mechanism that can reduce deviation and improve processing accuracy.

[0006] To achieve the above-mentioned objectives, the present invention provides the following technical solutions: a laser spot control system with an adaptive feedback mechanism, comprising: a workpiece detection module, a model generation module, a learning and analysis module, a working condition acquisition module, an optimization and adjustment module, and a hardware execution module; the workpiece detection module is used to collect workpiece spatial data and workpiece material composition; the model generation module is used to generate a digital twin based on the workpiece spatial data and workpiece material composition; the learning and analysis module is used to analyze the digital twin during the processing process and cooperate with the self-learning processing library to generate optimized process parameters, working paths, and to-be-processed data, and, after the processing is completed, generate optimized parameters and optimized paths based on the self-learning processing library; the working condition acquisition module is used to collect actual processing data and external environment data during the workpiece processing process; the optimization and adjustment module is used to generate compensation data and adjustment data based on external environment data, actual processing data, and to-be-processed data; the hardware execution module is used to process the workpiece according to the process parameters and working path.

[0007] By adopting the above technical solutions, the establishment of a digital twin can quickly find the optimal process parameters and working path before the workpiece is processed, reducing the trial and error costs, which is conducive to improving the efficiency and accuracy of workpiece processing.

[0008] The present invention is further configured as follows: the workpiece spatial data includes workpiece size, workpiece shape, and workpiece surface roughness; and the workpiece material composition includes element composition, thermal properties, and optical properties.

[0009] The present invention is further configured as follows: the learning analysis module includes a preset unit, a pseudo-processing unit and a self-learning unit; the preset unit is used to extract the process parameters and working paths corresponding to the digital twin from the self-learning processing library; the pseudo-processing unit is used to perform simulated processing on the digital twin according to the process parameters and working paths to generate pseudo-processing data; the self-learning unit is used to generate optimization parameters and optimization paths according to the process parameters and working paths corresponding to the digital twin in the self-learning processing library.

[0010] By adopting the above technical solution, the self-learning processing library is set in the deep reinforcement learning model of the embedded AI chip, and the results of the embedded AI learning are formed into a self-learning processing library. When the same workpiece is encountered again during the processing process, the data in the self-learning processing library can be called in time.

[0011] The present invention is further configured as follows: the optimization and adjustment module includes a compensation unit and an optimization unit; the compensation unit is used to generate compensation data based on the combined error and control law of external environment data, actual processing data, simulated processing data and stability algorithm; the optimization unit is used to generate adjustment data based on the combined error, control law, external environment data, actual processing data, simulated processing data and stability algorithm.

[0012] By adopting the above technical solution, the optimization and adjustment module is specifically an embedded AI chip deep reinforcement learning model. The deep reinforcement learning model run by the embedded AI chip continuously learns through trial and error through interaction with the environment to optimize the control strategy of the laser spot;

[0013] The present invention is further configured as follows: the hardware execution module includes an energy control unit and a position control unit; the energy control unit controls the laser output according to the process parameters; and the position control unit controls the laser position according to the working path.

[0014] The laser spot control method with an adaptive feedback mechanism uses the laser spot control system with an adaptive feedback mechanism as described above, and includes the following steps:

[0015] S1, collect workpiece spatial data and workpiece material composition;

[0016] S2, generating a digital twin based on the workpiece spatial data and workpiece material composition;

[0017] S3, analyze the digital twin and generate process parameters, work paths and intended processing data in conjunction with the self-learning processing library;

[0018] S4, processing the workpiece according to the process parameters and working path;

[0019] S5, collecting actual processing data and external environment data during the workpiece processing;

[0020] S6. Generate compensation data and adjustment data based on external environment data, actual processing data, proposed processing data and stability algorithm;

[0021] S7, optimize the process parameters according to the adjustment data, optimize the work path according to the compensation data, save them into the self-learning processing library, and jump to S4 until the processing is completed and jump to S8;

[0022] S8. Generate optimization parameters and optimization paths based on the process parameters and working paths of the corresponding digital twin in the self-learning processing library.

[0023] The present invention is further configured as follows: in S6, the compensation data and the adjustment data are calculated by an adaptive control algorithm of Lyapunov stability, and the adaptive control algorithm of Lyapunov stability has a joint error, a control law and a Lyapunov function.

[0024] The present invention is further configured as follows:

[0025] S61, using the external environment data, the actual processing data, and the intended processing data as input to define a joint error, and deriving compensation data based on the control law;

[0026] S62. Establish a Lyapunov function based on the joint error, control law, external environment data, actual processing data and intended processing data, and take its derivative to obtain the adjustment data.

[0027] By employing the aforementioned technical solution, the adaptive control algorithm for Lyapunov stability is primarily used to ensure the stability and rapid convergence of laser spot control systems with adaptive feedback mechanisms. This allows the system to automatically adjust compensation and adjustment data to maintain the precise shape and position of the laser spot in the face of environmental changes and fluctuating processing conditions. Compared to conventional algorithms, this algorithm, based on rigorous mathematical stability theory, can more effectively handle system uncertainties and nonlinearities, thereby reducing errors and providing more reliable and efficient control performance.

[0028] The present invention is further configured as follows: in S3, the process parameters include laser configuration parameters, laser operation parameters and cooling parameters.

[0029] The present invention is further configured as follows:

[0030] S81, generating a process parameter table based on time for the laser configuration parameters, laser operation parameters, and cooling parameters in the process parameters corresponding to the digital twin in the self-learning processing library;

[0031] S82, comparing the corresponding parameters in the process parameter table to see whether they change over time;

[0032] If it does not change over time, the corresponding process parameters in S3 are retained;

[0033] If it changes over time, generate a formula for the corresponding parameter;

[0034] S83, summarizing the corresponding process parameters and parameter formulas in S82 to generate optimized parameters;

[0035] S84. Remove external environment data based on the working path of the corresponding digital twin in the self-learning processing library to generate self-deviation data;

[0036] S85. Generate an optimized path based on its own deviation data and the working path in S3.

[0037] In summary, this application includes at least one of the following beneficial technical effects:

[0038] 1. By establishing a digital twin, the optimal process parameters and working path can be quickly found before the workpiece is processed, reducing the cost of trial and error, which is conducive to improving the efficiency and accuracy of workpiece processing.

[0039] 2. Through continuous trial and error learning through the interaction between the optimization and adjustment module and the environment, the control strategy of the laser spot is optimized; the learning results are formed into a self-learning processing library. When the same workpiece is encountered again during the processing process, the data in the self-learning processing library can be called in time.

[0040] 3. The adaptive control algorithm based on Lyapunov stability ensures the stability and rapid convergence of the system. When faced with environmental changes and fluctuations in processing conditions, it can automatically adjust compensation data and adjustment data to maintain the precise shape and position of the laser spot, reduce errors, and provide more reliable and efficient control performance. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 Schematic diagram of the relationship between the laser spot control system with adaptive feedback mechanism in the first embodiment of the present invention;

[0042] Figure 2 Schematic diagram of the relationship between signal processing modules in the third embodiment of the present invention;

[0043] Figure 3 This is a schematic diagram of the relationship between the fault prediction and health management modules in the third embodiment of the present invention;

[0044] Figure 4 Schematic diagram of steps S1 to S5 of the laser spot control method with an adaptive feedback mechanism in the second embodiment of the present invention;

[0045] Figure 5 Schematic diagram of steps S5 to S8 of the laser spot control method with an adaptive feedback mechanism in the second embodiment of the present invention; DETAILED DESCRIPTION

[0046] It should be noted that, in the absence of conflict, the embodiments and features of the embodiments in this application can be combined with each other. The present invention will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0047] It should be noted that, unless otherwise specified, all technical and scientific terms used in this application have the same meaning as commonly understood by ordinary technicians in the technical field to which this application belongs.

[0048] See also Figure 1-5 , the present invention provides the following technical solutions:

[0049] Example 1, see Figure 1 , a laser spot control system with an adaptive feedback mechanism, including a workpiece detection module, a model generation module, a learning analysis module, a working condition acquisition module, an optimization and adjustment module, and a hardware execution module;

[0050] The workpiece detection module is used to collect workpiece spatial data and workpiece material composition. The workpiece detection module mainly uses a high-speed CMOS camera and a material property sensor. The high-speed CMOS camera is a high-performance imaging device based on a complementary metal oxide semiconductor (CMOS) image sensor. In this system, a frame rate of 5000fps and a pixel size of 3.45μm×3.45μm are preferred. With a frame rate of 5000fps, subtle changes in the workpiece surface can be quickly captured in real time. For example, during subsequent workpiece processing, deformation or thermal effects on the workpiece surface can be monitored in real time to ensure timely feedback and adjustment of the laser spot. The high-resolution camera with a pixel size of 3.45μm×3.45μm can detect tiny surface defects or spot offsets, and can carefully monitor the contact between the laser and the workpiece surface to ensure processing accuracy. The material property sensor is a sensor based on LIBS spectral analysis that can detect the material properties of the workpiece element composition in real time. The material property sensor can quickly and accurately identify the workpiece material type and composition before laser processing.

[0051] Among them, the workpiece spatial data in the workpiece detection module includes workpiece size, workpiece shape, and workpiece surface roughness; the workpiece material composition in the workpiece detection module includes element composition, thermal properties, and optical properties.

[0052] The model generation module is used to generate a digital twin based on the workpiece spatial data and workpiece material composition;

[0053] The learning analysis module is used to analyze the digital twin during the processing process and generate optimized process parameters, work paths and intended processing data in conjunction with the self-learning processing library. After the processing is completed, the module also generates optimized parameters and optimized paths based on the self-learning processing library.

[0054] The learning analysis module includes a preset unit, a proposed processing unit, and a self-learning unit;

[0055] The preset unit is used to extract the process parameters and working paths corresponding to the digital twin from the self-learning processing library;

[0056] The pseudo-machining unit is used to simulate the machining of the digital twin according to the process parameters and work path to generate pseudo-machining data;

[0057] The self-learning unit is used to generate optimization parameters and optimization paths based on the process parameters and working paths of the corresponding digital twin in the self-learning processing library; among them, the self-learning processing library is formed by summarizing the results of embedded AI learning. When the same workpiece is encountered again during the processing process, the data in the self-learning processing library can be called in time.

[0058] The working condition acquisition module is used to collect actual processing data and external environment data during the workpiece processing process. The actual processing data is detected by laser wavefront sensors, high-speed CMOS cameras and other devices. The laser wavefront sensor is mainly used to detect the wavefront aberration of the laser beam. By monitoring the shape and phase changes of the wavefront, the system can understand the quality and focusing state of the beam in real time, thereby facilitating the subsequent adjustment of the optical system to ensure that the shape and energy distribution of the laser spot meet the processing requirements; the external environment data is collected by six-axis vibration sensors, infrared thermal imagers and other devices to detect data in the external environment that may affect the laser operation.

[0059] The optimization and adjustment module is used to generate compensation data and adjustment data based on external environment data, actual processing data, and planned processing data. The optimization and adjustment module is an embedded AI chip deep reinforcement learning model. The network structure of the embedded AI chip is a 3-layer LSTM + 2-layer CNN. The deep reinforcement learning model run by the embedded AI chip continuously learns through trial and error through interaction with the environment to optimize the laser spot control strategy. To reduce learning costs, strategies such as transfer learning, combining expert knowledge and experience, and optimizing algorithm structure and parameters can also be adopted;

[0060] The optimization and adjustment module includes a compensation unit and an optimization unit;

[0061] A compensation unit is used to generate compensation data based on the external environment data, actual processing data, simulated processing data and the combined error and control law of the stability algorithm;

[0062] The optimization unit is used to generate adjustment data according to the joint error, control law, external environment data, actual processing data, simulated processing data and stability algorithm.

[0063] The hardware execution module is used to process the workpiece according to the process parameters and work path. The hardware execution module mainly includes the voice coil motor drive platform, piezoelectric nanopositioning stage, water-cooled galvanometer system and composite zoom structure; the piezoelectric nanopositioning stage is mainly used for high-precision positioning and adjustment. It can fix the workpiece or optical components and achieve nanometer-level fine-tuning.

[0064] The hardware execution module includes an energy control unit and a position control unit;

[0065] The energy control unit controls the laser output according to the process parameters. The energy control unit includes a water-cooled galvanometer system and a compound zoom structure. The water-cooled galvanometer system combines a λ / 4 wave plate with a Glan prism to achieve 0-100% continuously adjustable laser energy attenuation control, which can improve the adjustment efficiency. The compound zoom structure dynamically adjusts the focusing ability of the laser beam.

[0066] The position control unit controls the laser position according to the working path; the position control unit includes a voice coil motor drive platform, a controller, etc., which is mainly used to control the rapid and precise movement of the laser spot to ensure that the laser spot is processed according to the preset path and speed.

[0067] Example 2, a laser spot control method with an adaptive feedback mechanism, uses the laser spot control system with an adaptive feedback mechanism, see Figure 4-5 , including the following steps:

[0068] S1, collect workpiece spatial data and workpiece material composition through high-speed CMOS camera and material property sensor;

[0069] S2. Modeling is performed based on the spatial data of the workpiece. A digital twin is generated through the workpiece material composition and modeling, i.e. a digital model with the same shape and material as the workpiece;

[0070] S3. Analyze the shape, material, and welding or cutting position of the digital twin, search for the closest or even consistent model from the self-learning processing library, and output process parameters, work path, and intended processing data; among them, process parameters include laser configuration parameters, laser operation parameters, and cooling parameters, namely welding current, voltage, speed, laser power, spot shape, cooling airflow parameters, weld spot diameter, etc.

[0071] By establishing a digital twin, the optimal process parameters and working path can be quickly found before the workpiece is processed, reducing the cost of trial and error, which is conducive to improving the efficiency and accuracy of workpiece processing.

[0072] S4. Configure the hardware execution module according to the process parameters, set the working path for the controller in the hardware execution module, and process the workpiece;

[0073] S5, collecting actual processing data and external environment data during the workpiece processing;

[0074] S6. Generate compensation data and adjustment data based on external environment data, actual processing data, proposed processing data and stability algorithm;

[0075] The specific steps of S6 are as follows:

[0076] S61, using the external environment data, the actual processing data, and the intended processing data as input to define a joint error, and deriving compensation data based on the control law;

[0077] S62. Establish a Lyapunov function based on the joint error, control law, external environment data, actual processing data, and intended processing data, and derive the function to obtain adjustment data;

[0078] For example, suppose the actual processing data is y(t), the working path in the simulated processing data is r(t), and the external environment data includes external disturbance d(t) and parameter uncertainty θ;

[0079] System Model:

[0080]

[0081] Among them, y is the actual processing data, is the regression vector, μ is the process parameter of the simulated processing data;

[0082] Define the joint error e(t):

[0083] e(t)=y(t)-r(t)-Δ(t)

[0084] Among them, △(t) is the expected deviation adjustment term obtained by integrating the external environment data d(t);

[0085] Design control law μ(t):

[0086]

[0087] Create a Lyapunov function:

[0088]

[0089] in, γ1, γ2 and γ3 are all greater than 0 and are adaptive gains;

[0090] Derivative V in the Lyapunov function established above, substitute the control law and error dynamics, and satisfy Get the adaptive law;

[0091] In this algorithm To compensate the data, the joint error e(t) is directly updated to reflect the environmental disturbance, and Δ(t) is the adjustment data used to correct the working path in the intended processing data;

[0092] This algorithm is primarily used to ensure the stability and rapid convergence of laser spot control systems with adaptive feedback mechanisms. This allows the system to automatically adjust compensation and adjustment data to maintain the precise shape and position of the laser spot in the face of environmental changes and fluctuating processing conditions. Compared to conventional algorithms, this algorithm, based on rigorous mathematical stability theory, can more effectively handle system uncertainties and nonlinearities, thereby reducing errors and providing more reliable and efficient control performance.

[0093] S7, optimize the process parameters according to the adjustment data, optimize the work path according to the compensation data, save them into the self-learning processing library, and jump to S4 until the processing is completed and jump to S8;

[0094] S8. Generate optimization parameters and optimization paths based on the process parameters and working paths of the corresponding digital twin in the self-learning processing library;

[0095] The specific steps of S8 are as follows:

[0096] S81, generating a process parameter table based on time for the laser configuration parameters, laser operation parameters, and cooling parameters in the process parameters corresponding to the digital twin in the self-learning processing library;

[0097] S82, comparing the corresponding parameters in the process parameter table to see whether they change over time;

[0098] If it does not change over time, the corresponding process parameters in S3 are retained;

[0099] If it changes over time, the corresponding parameter formula is generated;

[0100] S83, summarizing the corresponding process parameters and parameter formulas in S82 to generate optimized parameters;

[0101] S84. Remove external environment data based on the working path of the corresponding digital twin in the self-learning processing library to generate self-deviation data;

[0102] S85. Generate an optimized path based on its own deviation data and the working path in S3.

[0103] Example 3, see Figure 2-3 According to the laser spot control system with an adaptive feedback mechanism in the first embodiment, the system may further include:

[0104] The signal processing module uses a hybrid architecture processor based on an FPGA module. It is primarily responsible for processing underlying signals, including image signals from high-speed CMOS cameras, displacement signals from confocal displacement sensors, wavefront aberration signals from laser wavefront sensors, vibration signals from six-axis vibration sensors, temperature signals from infrared thermal imagers, and spectral signals from material property sensors. It processes a variety of signal types with a signal delay of less than 50μs, improving signal transmission efficiency and facilitating rapid action of the hardware execution module, further reducing the occurrence of deviations.

[0105] The fault prediction and health management module monitors the operating status of the laser spot control system with an adaptive feedback mechanism in real time, analyzes historical data and all sensor information, predicts potential faults, evaluates the health of the system, and reminds users to perform care and maintenance when necessary.

[0106] In the fourth embodiment, according to the laser spot control method with an adaptive feedback mechanism in the second embodiment, a multi-objective optimization algorithm is also used during processing;

[0107] The multi-objective optimization algorithm (NSGA-III) is used to balance processing speed, accuracy, and energy consumption. The multi-objective optimization algorithm (NSGA-III) is primarily used to balance the three mutually constrained objectives of processing speed, accuracy, and energy consumption during laser processing. Unlike conventional single-objective optimization algorithms, it can simultaneously address multiple objectives and find a set of Pareto optimal solutions.

[0108] Obviously, the embodiments described above are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts should fall within the scope of protection of the present invention.

Claims

1. A laser spot control system with an adaptive feedback mechanism, characterized by: include: Workpiece detection module, used to collect workpiece spatial data and workpiece material composition; A model generation module is used to generate a digital twin based on the workpiece spatial data and workpiece material composition; The learning and analysis module is used to analyze the digital twin during the processing and generate optimized process parameters, work paths, and intended processing data in conjunction with the self-learning processing library. After the processing is completed, the module also generates optimized parameters and optimized paths based on the self-learning processing library. Working condition acquisition module, used to collect actual processing data and external environment data during workpiece processing; Optimization and adjustment module, used to generate compensation data and adjustment data based on external environment data, actual processing data and planned processing data; as well as The hardware execution module is used to process the workpiece according to the process parameters and working path.

2. The laser spot control system with an adaptive feedback mechanism according to claim 1, characterized in that: The workpiece spatial data includes workpiece size, workpiece shape, and workpiece surface roughness; the workpiece material composition includes element composition, thermal properties, and optical properties.

3. The laser spot control system with an adaptive feedback mechanism according to claim 1, characterized in that: The learning analysis module includes a preset unit, a proposed processing unit and a self-learning unit; The preset unit is used to extract the process parameters and working paths corresponding to the digital twin from the self-learning processing library; The pseudo-machining unit is used to simulate machining on the digital twin according to the process parameters and the working path to generate pseudo-machining data; The self-learning unit is used to generate optimization parameters and optimization paths based on the process parameters and working paths of the corresponding digital twin in the self-learning processing library.

4. The laser spot control system with an adaptive feedback mechanism according to claim 2, characterized in that: The optimization and adjustment module includes a compensation unit and an optimization unit; The compensation unit is used to generate compensation data based on the external environment data, the actual processing data, the simulated processing data and the combined error and control law of the stability algorithm; The optimization unit is used to generate adjustment data according to the joint error, the control law, the external environment data, the actual processing data, the simulated processing data and the stability algorithm.

5. The laser spot control system with an adaptive feedback mechanism according to claim 3, characterized in that: The hardware execution module includes an energy control unit and a position control unit; The energy control unit is used to control the laser output according to the process parameters; The position control unit is used to control the laser position according to the working path.

6. A laser spot control method with an adaptive feedback mechanism, using the laser spot control system with an adaptive feedback mechanism as claimed in claim 4, characterized in that: The following steps are involved: S1, collect workpiece spatial data and workpiece material composition; S2, generating a digital twin based on the workpiece spatial data and workpiece material composition; S3, analyze the digital twin and generate process parameters, work paths and intended processing data in conjunction with the self-learning processing library; S4, processing the workpiece according to the process parameters and working path; S5, collecting actual processing data and external environment data during the workpiece processing; S6. Generate compensation data and adjustment data based on external environment data, actual processing data, proposed processing data and stability algorithm; S7, optimize the process parameters according to the adjustment data, optimize the work path according to the compensation data, save them into the self-learning processing library, and jump to S4 until the processing is completed and jump to S8; S8. Generate optimization parameters and optimization paths based on the process parameters and working paths of the corresponding digital twin in the self-learning processing library.

7. The laser spot control method with an adaptive feedback mechanism according to claim 5, wherein the compensation data and adjustment data in S6 are calculated by an adaptive control algorithm of Lyapunov stability, and the adaptive control algorithm of Lyapunov stability has a joint error, a control law and a Lyapunov function.

8. The laser spot control method with an adaptive feedback mechanism according to claim 6, wherein the adaptive control algorithm of Lyapunov stability comprises the following steps: S61, using the external environment data, the actual processing data, and the intended processing data as input to define a joint error, and deriving compensation data based on the control law; S62. Establish a Lyapunov function based on the joint error, control law, external environment data, actual processing data and intended processing data, and take its derivative to obtain the adjustment data.

9. The laser spot control method with an adaptive feedback mechanism according to claim 5, wherein the process parameters in S3 include: Laser configuration parameters, laser operating parameters and cooling parameters.

10. The laser spot control method with an adaptive feedback mechanism according to claim 8, wherein generating the optimization parameters and optimization path according to the process parameters and working path corresponding to the digital twin in the self-learning processing library in S8 comprises the following steps: S81, generating a process parameter table based on time for the laser configuration parameters, laser operation parameters, and cooling parameters in the process parameters corresponding to the digital twin in the self-learning processing library; S82, comparing the corresponding parameters in the process parameter table to see whether they change over time; If it does not change over time, the corresponding process parameters in S3 are retained; If it changes over time, generate a formula for the corresponding parameter; S83, summarizing the corresponding process parameters and parameter formulas in S82 to generate optimized parameters; S84. Remove external environment data based on the working path of the corresponding digital twin in the self-learning processing library to generate self-deviation data; S85. Generate an optimized path based on its own deviation data and the working path in S3.

Images