Laser processing method and device based on process state learning

Abstract

The present application relates to a laser processing method and device based on process state learning, wherein a photodiode is used to measure the pulse energy value of each laser pulse in the laser processing process in real time, the subtraction value of the pulse energy value and the target pulse energy output value is calculated to obtain the pulse energy error, and when the compensation condition is triggered in the reprocessing process at the defect position, the galvanometer speed is automatically adjusted and the overlap rate is automatically adjusted. Wherein, the galvanometer speed represents the galvanometer scanning speed, and the overlap rate represents the pulse overlap rate. When the compensation condition is triggered, the system automatically adjusts the galvanometer scanning speed and the pulse overlap rate to suppress the influence of single pulse energy fluctuation on the processing quality, thereby solving the point-by-point fluctuation problem of material removal amount and reducing the range of heat affected zone.

Description

Technical Field

[0001] This invention relates to the field of laser precision machining technology, and in particular to a laser machining method and apparatus based on process state learning. Background Technology

[0002] In semiconductor and display manufacturing processes, critical steps such as micro-patterning, glass cutting, thin film removal, and drilling commonly face challenges caused by the insufficient stability of ultrashort pulse lasers. Specifically, for ultrashort pulse lasers with pulse widths in the picosecond or femtosecond range, even minute energy differences or temporal waveform fluctuations between individual pulses are sufficient to cause significant deterioration in processing quality, ultimately leading to a severe drop in product yield. This instability of individual pulses directly causes point-to-point fluctuations in material removal, making it difficult to maintain uniform processing depth and width. Simultaneously, abnormal fluctuations in pulse energy can lead to uncontrolled thermal accumulation during laser-material interaction, significantly increasing the size of the heat-affected zone and inducing secondary defects such as microcracks, molten spatter, and thermal stress concentration at material edges. Summary of the Invention

[0003] This invention provides a laser processing method and apparatus based on process state learning, aiming to solve at least one of the technical problems existing in the prior art.

[0004] The technical solution of this invention is a laser processing method based on process state learning, comprising the following steps: S100: The pulse energy value of each laser pulse during laser processing is measured in real time using a photodiode. S200. Calculate the difference between the pulse energy value and the target pulse energy output value to obtain the pulse energy error; S300: During the reprocessing of the defect location, when the compensation condition is triggered, the galvanometer speed and overlap rate are automatically adjusted.

[0005] According to some embodiments of the present invention, in step S300, If the absolute value of the pulse energy error is greater than the energy error threshold, the compensation condition is triggered; If the cumulative defect index is greater than the cumulative threshold of the heat map, the compensation condition is triggered; If the number of abnormal laser pulses generated at the same location exceeds the repetition threshold, the compensation condition is triggered. If the quality degradation index exceeds the quality index threshold, the compensation condition is triggered.

[0006] According to some embodiments of the present invention, the adjustment of the galvanometer speed is expressed as follows: , In the formula, The adjusted galvanometer speed, As the reference galvanometer velocity, This is the compensation coefficient for the galvanometer velocity. This represents the strength of a local defect.

[0007] According to some embodiments of the present invention, the adjustment of the overlap rate is expressed as follows: , In the formula, The adjusted overlap rate. As the baseline overlap rate, This is the compensation coefficient for the overlap rate. The pulse energy error is denoted as .

[0008] According to some embodiments of the present invention, the method further includes the following steps: S400. Generate and store the defect heat map data accumulated in the previous process, wherein when anomalies occur repeatedly at an abnormal location, the heat map value of the abnormal location is increased. S500. When the abnormal position repeatedly experiences insufficient or excessive pulse energy, the laser output compensation value is increased or decreased in advance before entering the abnormal position to obtain the pre-compensated output power value. S600, Laser processing is performed at the abnormal location using the pre-compensated output power value.

[0009] According to some embodiments of the present invention, the heat map values ​​are represented as follows: , In the formula, for The cumulative defect index at coordinate location, i.e. The thermal map values ​​for the coordinates, where N represents the number of anomalous laser pulses generated at the same location. The weighting coefficient for the pulse energy error of the i-th laser pulse that experiences an anomaly at the same location. The pulse energy error is denoted as .

[0010] According to some embodiments of the present invention, in step S500, The laser output compensation value includes a real-time position compensation value and a historical position self-learning compensation value. The real-time position compensation value is obtained by multiplying the real-time pulse energy error value with the error feedback coefficient. The pre-compensated output power value is generated based on the reference output power value, the real-time position compensation value, and the historical position self-learning compensation value, as shown below: , In the formula, For coordinates The pre-compensated output power value at the output. The reference output power value, The error feedback coefficient is... The real-time pulse energy error value refers to the pulse energy error value obtained through real-time detection. The historical location self-learning compensation value is stored in the heatmap database. Long-term compensation value for location.

[0011] According to some embodiments of the present invention, the method further includes the following steps: S700: Store the location of the detected abnormal laser pulse as a defect dataset. After the process is completed, only the stored defect locations are selectively reprocessed. S800. Calculate the reprocessing priority of the defect location, and perform laser processing on the defect location according to the reprocessing priority.

[0012] According to some embodiments of the present invention, the reprocessing priority of the defect location is expressed as follows: , In the formula, for The priority of reprocessing at different locations. As the baseline overlap rate, for The cumulative defect index at coordinate location, For defect density, As an indicator of quality deterioration, This is the cumulative defect weighting coefficient, also known as the heatmap influence coefficient. This is the defect impact coefficient. This represents the coefficient of quality degradation impact.

[0013] The present invention also relates to a computer device, including a memory and a processor, wherein the processor executes the method described above when executing a computer program stored in the memory.

[0014] The present invention also relates to a computer-readable storage medium having program instructions stored thereon, which, when executed by a processor, implement the method described above.

[0015] The beneficial effects of this invention include: Real-time measurement of the pulse energy value of each laser pulse during laser processing using a photodiode; calculation of the difference between the pulse energy value and the target pulse energy output value to obtain the pulse energy error; and automatic adjustment of the galvanometer speed and overlap rate during reprocessing at the defect location when a compensation condition is triggered. Here, galvanometer speed represents the galvanometer scanning speed, and overlap rate represents the pulse overlap rate. When the compensation condition is triggered, the system automatically adjusts the galvanometer scanning speed and pulse overlap rate to suppress the impact of single-pulse energy fluctuations on processing quality, thereby solving the problem of point-to-point fluctuations in material removal and reducing the range of the heat-affected zone.

[0016] Furthermore, additional aspects and advantages of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Attached Figure Description

[0017] Figure 1 This is an optional flowchart of a laser processing method based on process state learning in an embodiment of the present invention.

[0018] Figure 2 This is a flowchart of real-time pulse processing in an embodiment of the present invention. Detailed Implementation

[0019] The following will provide a clear and complete description of the concept, specific structure, and technical effects of the present invention in conjunction with the embodiments and accompanying drawings, so as to fully understand the purpose, solution, and effects of the present invention. It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other.

[0020] It should be noted that, unless otherwise specified, when a feature is referred to as "fixed" or "connected" to another feature, it can be directly fixed or connected to the other feature, or indirectly fixed or connected to the other feature. Furthermore, the descriptions of "upper," "lower," "left," "right," "top," and "bottom" used in this invention are only relative to the relative positional relationships of the various components of the invention in the accompanying drawings.

[0021] Furthermore, unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art. The terminology used in this specification is for the purpose of describing particular embodiments only and not for limiting the invention. The term "and / or" as used herein includes any combination of one or more of the associated listed items.

[0022] It should be understood that although the terms first, second, third, etc., may be used to describe various elements in this invention, these elements should not be limited to these terms. These terms are only used to distinguish elements of the same type from each other. For example, a first element may also be referred to as a second element without departing from the scope of this invention, and similarly, a second element may also be referred to as a first element.

[0023] Reference Figures 1 to 2 In some embodiments, the laser processing method based on process state learning of the present invention includes at least the following steps: S100: The pulse energy value of each laser pulse during laser processing is measured in real time using a photodiode. S200. Calculate the difference between the pulse energy value and the target pulse energy output value to obtain the pulse energy error; S300: During the reprocessing of the defect location, when the compensation condition is triggered, the galvanometer speed and overlap rate are automatically adjusted.

[0024] As can be seen, the galvanometer speed represents the galvanometer scanning speed, and the overlap rate represents the pulse overlap rate. When the compensation condition is triggered, the system automatically adjusts the galvanometer scanning speed and the pulse overlap rate to suppress the influence of single-pulse energy fluctuations on the processing quality, thereby solving the problem of point-to-point fluctuations in material removal and reducing the range of the heat-affected zone.

[0025] Specifically, this invention also provides a laser processing system based on process state learning, including a laser pulse unit energy measurement unit, a galvanometer position synchronization unit, a real-time process analysis unit, a defect map database, and a selective reprocessing control unit. By comprehensively analyzing pulse energy information, spatial position information, processing quality information, and detection result information, it not only controls the laser output but also uniformly manages the following process parameters: galvanometer speed, overlap rate, laser burst conditions, fill spacing, defocusing amount, reprocessing path, spot size, polarization state, repeated processing direction, number of repeated processing cycles, and laser pulse repetition frequency. These parameters are all closely related to the overall process formulation. Through this unified management, the system itself can proactively perform maintenance, management, compensation, and reprocessing. In other words, this invention does not merely stabilize the output of existing processes but develops an active adaptive manufacturing system that integrates process data, position information, detection information, and reprocessing information, enabling the system to learn conditions, conduct comparative analysis, and autonomously determine whether countermeasures are needed.

[0026] In one specific embodiment, the adaptive output correction system: The system measures the energy value of each laser pulse unit in real time and automatically calculates the error between it and the target output.

[0027] When the error range exceeds the threshold, the control unit will not only adjust the output of the next pulse, but also change the galvanometer speed and overlap conditions at the same time.

[0028] Therefore, the device according to this embodiment can correct minute fluctuations in laser output in real time on a pulse unit.

[0029] In a specific embodiment, the pulse energy value is expressed as follows: , In the formula, Let be the pulse energy value of the i-th laser pulse. For machining position coordinates, The duration of the pulse. This is a function that calculates laser pulse energy by inputting the processing position and pulse duration.

[0030] In some embodiments, during step S300, If the absolute value of the pulse energy error is greater than the energy error threshold, the compensation condition is triggered. If the cumulative defect index is greater than the cumulative threshold of the heatmap, the compensation condition is triggered; If the number of abnormal laser pulses generated at the same location exceeds the repetition threshold, the compensation condition is triggered. If the quality deterioration index exceeds the quality index threshold, the compensation condition is triggered.

[0031] In some embodiments, adjusting the galvanometer speed is represented as follows: , In the formula, The adjusted galvanometer speed, As the reference galvanometer velocity, This is the compensation coefficient for the galvanometer velocity. This represents the strength of a local defect.

[0032] In some embodiments, the overlap rate is adjusted as follows: , In the formula, The adjusted overlap rate. As the baseline overlap rate, This is the compensation coefficient for the overlap rate. This refers to pulse energy error.

[0033] In some embodiments, the laser processing method based on process state learning further includes the following steps: S400: Generate and store the defect heat map data accumulated in the previous process, wherein when anomalies occur repeatedly at an abnormal location, the heat map value of the abnormal location is increased. S500 When insufficient or excessive pulse energy repeatedly occurs at abnormal positions, the laser output compensation value is increased or decreased in advance before entering the abnormal position to obtain the pre-compensated output power value. S600: Laser processing is performed at abnormal locations using the pre-compensated output power value.

[0034] It should be understood that using pre-compensated output power values ​​for laser processing at abnormal locations achieves forward-looking compensation, which helps to suppress the impact of single-pulse energy fluctuations on processing quality, thereby solving the problem of point-to-point fluctuations in material removal and reducing the range of the heat-affected zone.

[0035] In one specific embodiment, a location-based pre-compensation system: The system generates and stores defect heatmap data accumulated in previous processes.

[0036] When a specific location repeatedly experiences insufficient or excessive energy, the system pre-increases or decreases the laser output compensation value before entering that location.

[0037] That is, perform prediction-based pre-compensation before real-time feedback.

[0038] In some embodiments, the heatmap values ​​are represented as follows: , In the formula, for The cumulative defect index at coordinate location, i.e. The thermal map values ​​for the coordinates, where N represents the number of anomalous laser pulses generated at the same location. The weighting coefficient for the pulse energy error of the i-th laser pulse that experiences an anomaly at the same location. This refers to pulse energy error.

[0039] In some embodiments, during step S500, The laser output compensation value includes the real-time position compensation value and the historical position self-learning compensation value. The real-time position compensation value is obtained by multiplying the real-time pulse energy error value and the error feedback coefficient. Based on the baseline output power value, the real-time position compensation value, and the historical position self-learning compensation value, the pre-compensated output power value is generated, as shown below: , In the formula, For coordinates The pre-compensated output power value at the output. As the reference output power value, For error feedback coefficients, This refers to the real-time pulse energy error value, i.e., the pulse energy error value obtained through real-time detection. These are the self-learning compensation values ​​for historical locations, i.e., those stored in the heatmap database. Long-term compensation value for location.

[0040] Specifically, the real-time pulse energy error value is expressed as follows: , In the formula, This represents the real-time pulse energy error value. The target pulse energy output value, This represents the pulse energy value.

[0041] In this process, a portion of the laser extracted by the beam splitter is detected by a photodiode. This value is then converted by an ADC and converted into a pulse energy value using a pre-calibrated conversion coefficient. This pulse energy value is then compared with the target pulse energy output value to calculate the real-time pulse energy error value.

[0042] Specifically, if insufficient energy persists at the same location after the process is completed, a long-term compensation value is increased. This is repeated as the process is carried out. It will gradually learn, thereby reducing compensation errors or using the learned data to improve process characteristics.

[0043] To define the limiting condition for energy compensation, that is, the limit value that laser can compensate for, the formula is... Revised as follows: , In the formula, For coordinates The pre-compensated output power value at the output. As the reference output power value, For error feedback coefficients, This refers to the real-time pulse energy error value, i.e., the pulse energy error value obtained through real-time detection. These are the self-learning compensation values ​​for historical locations, i.e., those stored in the heatmap database. The long-term compensation value for the location, where Clamp is the limiting function. This represents the baseline amplitude of the self-learning compensation value for historical positions. This indicates the safe upper limit of the output power value after pre-compensation.

[0044] Specifically, the long-term compensation values ​​stored in the heatmap database can be accumulated and averaged by measuring energy errors at specific locations (x, y), and calculated and updated according to the following formula: , In the formula, Indicates the updated Long-term compensation value of location, Indicates the state before the update The long-term compensation value of a location, i.e., the current long-term compensation value of the current location. Indicates the learning rate. For error feedback coefficients, This represents the real-time pulse energy error value.

[0045] Among them, the updated The long-term compensation value of the location is written to the heatmap database to replace the original data.

[0046] In some embodiments, the laser processing method based on process state learning further includes the following steps: S700: Store the location of the detected abnormal laser pulse as a defect dataset. After the process is completed, only the stored defect locations are selectively reprocessed. S800: Calculate the reprocessing priority of the defect location, and perform laser processing on the defect location according to the reprocessing priority.

[0047] In one specific embodiment, the automated reprocessing system: The system stores the locations of detected abnormal pulses as a defect dataset.

[0048] After the process is completed, the system selectively rescans the stored defect locations, and the reprocessing conditions are set independently from the existing process conditions.

[0049] This allows for a reduction in overall reprocessing time while increasing production yield.

[0050] In some embodiments, the reprocessing priority of defect locations is represented as follows: , In the formula, for The priority of reprocessing at different locations. As the baseline overlap rate, for The cumulative defect index at coordinate location, For defect density, As an indicator of quality deterioration, This is the cumulative defect weighting coefficient, also known as the heatmap influence coefficient. This is the defect impact coefficient. This represents the coefficient of quality degradation impact.

[0051] In a specific embodiment, the laser processing method based on process state learning further includes the following steps: based on the information accumulated through self-learning, for each batch of each material, select a process formula that matches each material for processing.

[0052] In one specific embodiment, a self-learning system linked to quality inspection: The system, through its self-learning function and in conjunction with quality inspection, can proactively identify the causes of previous defects, thereby minimizing the defect rate in the next batch.

[0053] In a specific embodiment, the laser processing method based on process state learning further includes the following steps: transmitting the reprocessing results back to the heat map generation module and updating the self-learning database.

[0054] Specifically, by integrating the energy information, position information, process history information, and quality inspection information of laser pulse units, process conditions are compensated in real time and selective reprocessing is performed.

[0055] First, a space-based energy distribution model A portion of the laser output is measured via a photodiode for each laser pulse i: , in, This function represents the calculation of the laser pulse energy at a given laser irradiation location, laser irradiation duration, optical system correction coefficients, and device status. This function can be implemented using at least one of the following methods: sensor measurements, lookup tables (LUTs), regression models, or artificial intelligence prediction models.

[0056] The energy information calculated here, Figure 2 In the step of galvanometer position synchronization, the position information is synchronized, and in Figure 2 In the spatial coordinate mapping process, it is associated with the position coordinates.

[0057] Second, the heatmap of cumulative location error. , in, The cumulative defect index represents the cumulative defect index at different locations, i.e. The thermal map values ​​for the coordinates, where N represents the number of anomalous laser pulses generated at the same location. The weighting coefficient for the pulse energy error of the i-th laser pulse that experiences an anomaly at the same location. This refers to pulse energy error.

[0058] When anomalies repeatedly occur at a specific location (i.e., the aforementioned abnormal location), the heatmap values ​​increase accordingly. The calculated results are stored in a heatmap database to represent the recurring anomalies at the specific location in a spatially cumulative manner, and are used in the subsequent defect area determination step for identifying risk factors.

[0059] Third, the position adaptive compensation formula Based on the heat map database and past process history, long-term compensation values ​​are generated for each location.

[0060] , in, For the laser control unit in coordinate The pre-compensated output power value at the output. This is the baseline output power value set within the process formulation. The error feedback coefficient is calculated in real time. This refers to the real-time pulse energy error value, i.e., the pulse energy error value obtained through real-time detection. These are the self-learning compensation values ​​for historical locations, i.e., those stored in the heatmap database. Long-term compensation value for location.

[0061] Before entering a specific location, abnormal coordinates are actively identified and pre-compensated. The calculated results are transmitted to the laser control unit, and... Figure 2 This is performed during the laser output compensation step.

[0062] Fourth, adaptive control formula for process conditions The galvanometer speed is automatically adjusted as the degree of defect increases.

[0063] , Applied to Figure 2 In the process parameter adjustment steps; in, The adjusted galvanometer speed, As the reference galvanometer velocity, This is the compensation coefficient for the galvanometer speed set in the process formulation. This represents the strength of a local defect.

[0064] When a local defect occurs, its location is determined, and parameters such as the galvanometer speed are automatically adjusted according to compensation conditions. As the defect increases, the galvanometer speed automatically decreases. The calculated results are transmitted to the galvanometer speed controller for further adjustment.

[0065] The compensation condition we want to express here is: a coefficient related to the process formulation, that is, an optical system spatial compensation coefficient that includes factors such as laser output, galvanometer field of view characteristics, and spot size variation; where... This is a proportionality coefficient used to determine the intensity of local defects. Determine the adjustment range for the galvanometer speed.

[0066] Fifth, adaptive overlap rate The overlap rate condition is dynamically changed based on the pulse energy error.

[0067] Applied to Figure 2 In the process parameter adjustment steps.

[0068] , in, The adjusted overlap rate. As the baseline overlap rate, The compensation coefficient is used to determine the extent of the increase in overlap rate. This refers to pulse energy error.

[0069] The overlap rate condition is also automatically adjusted according to the compensation condition.

[0070] The system will perform a compensation action when any one or more of the following conditions are met: The absolute value of the pulse energy error is greater than the energy error threshold; The cumulative defect index is greater than the cumulative threshold of the heatmap; The number of abnormal laser pulses generated at the same location exceeds the repetition threshold. The quality deterioration index is greater than the quality index threshold.

[0071] For example, if the allowable ranges for each indicator are set as shown in Table 1: Based on various threshold standards, the system compensates for parameters such as galvanometer speed, laser output power, and pulse overlap rate in the next process according to priority, and replaces them with parameters that meet the standards before putting them into processing.

[0072] If the overlap rate is set as the priority adjustment item, the scan line spacing will be automatically calculated using the following formula: , In the formula, D represents the size of the light spot after focusing through the lens. This refers to the scan line spacing. The adjusted overlap rate ultimately forms the following control logic: Increase overlap rate --> Reduce scan line spacing --> Increase energy density, thereby reducing the number of defects and transmitting the data to the galvanometer and laser control unit.

[0073] Table 1. Permissible ranges for each indicator

[0074] Sixth, reprocessing priority function Using defect maps and quality information, calculate reprocessing priorities.

[0075] , The results calculated here will be used to generate a rework priority list and passed to the selective rework scheduling module.

[0076] Seventh, Process Recipe Manager Based on the information accumulated through self-learning, the system can select a suitable formula for processing each material (each batch) based on its material information.

[0077] It is understood that the present invention provides an adaptive production system capable of analyzing real-time process status and dynamically changing the overall processing conditions of the equipment.

[0078] In another possible implementation, laser processing methods based on process state learning include: Real-time acquisition of the pulse energy value of each laser pulse. The location information of the pulse energy value of each laser pulse is mapped to the grid cells of the XY coordinate system to determine the spatial area illuminated by each laser pulse. Laser pulses falling within the same grid are grouped to form a spatially based data set. The pulse energy error is accumulated within each grid to calculate the heatmap value; Grid areas with heatmap values ​​exceeding the cumulative threshold are identified as defect areas. Based on the magnitude of the heat map values, the magnitude of the pulse energy error, and the number of abnormal laser pulses, the defect area is classified into low-risk, medium-risk, or high-risk levels. When there are multiple defective areas, the rework priority of the defective areas is calculated by considering the size of the defective area, whether the defective area is a critical area, and the production efficiency. The defective areas are then laser-processed according to the rework priority.

[0079] Specifically, during laser processing, the pulse energy error generated pulse by pulse is spatially accumulated with the position information to generate a heat map.

[0080] Specifically, pulse error data acquisition: The laser control unit collects the pulse energy value, laser irradiation position coordinates, laser irradiation duration, galvanometer coordinates, and quality inspection results for each laser pulse in real time.

[0081] Location mapping: The acquired location information is mapped to grid cells in the XY coordinate system to determine the spatial region illuminated by each pulse.

[0082] Grid Classification: Pulses falling within the same grid are grouped to form a data set in spatial units.

[0083] Error accumulation: Within each grid, data such as energy deviation, location error, output fluctuation rate, and quality inspection results are accumulated to calculate the cumulative error for each region. Heatmap values ​​can be calculated using the following formula: .

[0084] Heatmap generation: A heatmap covering the entire processing area is generated using the accumulated error data. The heatmap visually displays the spatial distribution of the probability of defects occurring at various locations.

[0085] Defect area identification: Areas whose heatmap values ​​exceed the cumulative threshold are identified as defect areas.

[0086] Risk level classification: Based on factors such as the magnitude of heat map values, cumulative error, and frequency of recurrence, defective areas are classified into three risk levels: low, medium, and high.

[0087] Priority sorting: When there are multiple defective areas, the rework priority is calculated; the priority is determined by a combination of factors such as defect area, whether it is a critical area, and production efficiency.

[0088] Selective rework areas: Ultimately, only the areas requiring repair are extracted and transmitted to subsequent compensation control and repair processes.

[0089] Specifically, the defect information calculated from the thermal map is used to compensate for the laser output power and galvanometer control parameters in real time. Read the cumulative defect index corresponding to the current coordinate from the heat map; The system calculates the power compensation, galvanometer velocity compensation, and overlap rate compensation respectively. Among them, the historical position self-learning compensation value is the optical system spatial correction coefficient, which covers influencing factors such as F-theta lens distortion, galvanometer field characteristics, and spot size changes; Among them, the higher the thermal map value, the lower the galvanometer running speed, thereby increasing the energy density at the same location; The greater the energy deviation, the greater the pulse overlap rate, ensuring uniform and consistent processing quality. The calculated compensation parameters are sent to the galvanometer controller and the laser controller to perform selective rework. The rework results are sent back to the heat map generation module to update the self-learning database; Based on this closed-loop architecture, process parameters can be iteratively optimized repeatedly.

[0090] This invention also provides a computer device including a memory and a processor, wherein the processor performs the above-described method when executing a computer program stored in the memory.

[0091] This invention also provides a computer-readable storage medium storing program instructions thereon, which, when executed by a processor, implement the method described above.

[0092] It should be understood that the method steps in the embodiments of the present invention can be implemented or carried out by computer hardware, a combination of hardware and software, or by computer instructions stored in a non-transitory computer-readable storage medium. The method can use standard programming techniques. Each program can be implemented in a high-level procedural or object-oriented programming language to communicate with the computer system. However, if necessary, the program can be implemented in assembly or machine language. In any case, the language can be a compiled or interpreted language. Furthermore, for this purpose, the program can run on a programmed application-specific integrated circuit (ASIC).

[0093] Furthermore, the procedures described herein may be performed in any suitable order unless otherwise indicated herein or otherwise clearly contradicted by the context. The procedures described herein (or variations and / or combinations thereof) may be executed under the control of one or more computer systems configured with executable instructions, and may be implemented by hardware or a combination thereof as code (e.g., executable instructions, one or more computer programs, or one or more applications) that commonly executes on one or more processors. The computer program comprises a plurality of instructions executable by one or more processors.

[0094] Furthermore, the method can be implemented in any suitable type of computing platform, including but not limited to personal computers, minicomputers, mainframes, workstations, networked or distributed computing environments, standalone or integrated computer platforms, or in communication with charged particle tools or other imaging devices, etc. Aspects of the invention can be implemented as machine-readable code stored on a non-transitory storage medium or device, whether removable or integrated into a computing platform, such as a hard disk, optical read and / or write storage medium, RAM, ROM, etc., such that it is readable by a programmable computer, and when the storage medium or device is read by the computer, it can be used to configure and operate the computer to perform the processes described herein. Furthermore, the machine-readable code, or portions thereof, can be transmitted via wired or wireless networks. The invention described herein includes these and other different types of non-transitory computer-readable storage media when such media comprises instructions or programs that implement the steps described above in conjunction with a microprocessor or other data processor. When programmed according to the methods and techniques described in the invention, the invention may also include the computer itself.

[0095] A computer program can be applied to input data to perform the functions described herein, thereby transforming the input data to generate output data stored in non-volatile memory. The output information can also be applied to one or more output devices, such as a display. In a preferred embodiment of the invention, the transformed data represents physical and tangible objects, including specific visual depictions of physical and tangible objects generated on the display.

[0096] The above description is merely a preferred embodiment of the present invention. The present invention is not limited to the above-described embodiments. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention, as long as they achieve the technical effects of the present invention by the same means, should be included within the scope of protection of the present invention. Within the scope of protection of the present invention, the technical solutions and / or implementation methods can have various modifications and variations.

Claims

1. A laser processing method based on process state learning, characterized in that, Includes the following steps: S100: The pulse energy value of each laser pulse during laser processing is measured in real time using a photodiode. S200. Calculate the difference between the pulse energy value and the target pulse energy output value to obtain the pulse energy error; S300: During the reprocessing of the defect location, when the compensation condition is triggered, the galvanometer speed and overlap rate are automatically adjusted.

2. The laser processing method based on process state learning according to claim 1, characterized in that, In step S300 If the absolute value of the pulse energy error is greater than the energy error threshold, the compensation condition is triggered; If the cumulative defect index is greater than the cumulative threshold of the heat map, the compensation condition is triggered; If the number of abnormal laser pulses generated at the same location exceeds the repetition threshold, the compensation condition is triggered. If the quality degradation index exceeds the quality index threshold, the compensation condition is triggered.

3. The laser processing method based on process state learning according to claim 1, characterized in that, The adjustment of the galvanometer speed is expressed as follows: ， In the formula, The adjusted galvanometer speed, As the reference galvanometer velocity, This is the compensation coefficient for the galvanometer velocity. This represents the strength of a local defect.

4. The laser processing method based on process state learning according to claim 1, characterized in that, The adjustment of the overlap rate is expressed as follows: ， In the formula, The adjusted overlap rate. As the baseline overlap rate, This is the compensation coefficient for the overlap rate. The pulse energy error is denoted as .

5. The laser processing method based on process state learning according to claim 1, characterized in that, It also includes the following steps: S400. Generate and store the defect heat map data accumulated in the previous process, wherein when anomalies occur repeatedly at an abnormal location, the heat map value of the abnormal location is increased. S500. When the abnormal position repeatedly experiences insufficient or excessive pulse energy, the laser output compensation value is increased or decreased in advance before entering the abnormal position to obtain the pre-compensated output power value. S600, Laser processing is performed at the abnormal location using the pre-compensated output power value.

6. The laser processing method based on process state learning according to claim 5, characterized in that, The heatmap values ​​are represented as follows: ， In the formula, for The cumulative defect index at coordinate location, i.e. The thermal map values ​​for the coordinates, where N represents the number of anomalous laser pulses generated at the same location. The weighting coefficient for the pulse energy error of the i-th laser pulse that experiences an anomaly at the same location. The pulse energy error is denoted as .

7. The laser processing method based on process state learning according to claim 5, characterized in that, In step S500 The laser output compensation value includes a real-time position compensation value and a historical position self-learning compensation value. The real-time position compensation value is obtained by multiplying the real-time pulse energy error value with the error feedback coefficient. The pre-compensated output power value is generated based on the reference output power value, the real-time position compensation value, and the historical position self-learning compensation value, as shown below: ， In the formula, For coordinates The pre-compensated output power value at the output. The reference output power value, The error feedback coefficient is... The real-time pulse energy error value refers to the pulse energy error value obtained through real-time detection. The historical location self-learning compensation value is stored in the heatmap database. Long-term compensation value for location.

8. The laser processing method based on process state learning according to claim 1, characterized in that, It also includes the following steps: S700: Store the location of the detected abnormal laser pulse as a defect dataset. After the process is completed, only the stored defect locations are selectively reprocessed. S800. Calculate the reprocessing priority of the defect location, and perform laser processing on the defect location according to the reprocessing priority.

9. The laser processing method based on process state learning according to claim 8, characterized in that, The reprocessing priority of the defect location is represented as follows: ， In the formula, for The priority of reprocessing at different locations. As the baseline overlap rate, for The cumulative defect index at coordinate location, For defect density, As an indicator of quality deterioration, This is the cumulative defect weighting coefficient, also known as the heatmap influence coefficient. This is the defect impact coefficient. This represents the coefficient of quality degradation impact.

10. A computer device comprising a memory and a processor, characterized in that, When the processor executes a computer program stored in the memory, it performs the method as described in any one of claims 1 to 9.

Images