An Adaptive Precision Control Method and System for Femtosecond Laser Welding of Photovoltaic Glass

Abstract

This invention belongs to the interdisciplinary field of laser processing and photovoltaic manufacturing. It discloses an adaptive precision control method and system for femtosecond laser welding of photovoltaic glass. Through core technologies such as laser parameter compensation based on the characteristics of photovoltaic glass welding environment interference, collaborative synchronous control of multiple weld seams of femtosecond laser, and real-time evaluation of photovoltaic glass welding quality based on parameter feedback optimization, it breaks through the technical bottlenecks of traditional welding technology, such as "weak anti-interference ability, poor synchronization of multiple weld seams, and lagging quality control". It achieves real-time adaptive compensation for welding environment interference, micron-level synchronous control of multi-weld seam welding, and online closed-loop optimization of welding quality. It significantly improves the environmental adaptability of photovoltaic glass welding and the consistency of multi-weld seam formation, and is suitable for the high-precision welding requirements of various types of photovoltaic glass such as ultra-white, tempered, and flexible photovoltaic glass in mass production scenarios.

Description

Technical Field

[0001] This invention belongs to the field of information technology, specifically relating to an adaptive precision control method and system for femtosecond laser welding of photovoltaic glass. Background Technology

[0002] In the current field of femtosecond laser welding of photovoltaic glass, existing technologies have solved the core problems of "small heat-affected zone and basic weld formation," achieving initial industrial applications of photovoltaic glass welding. However, in practical engineering projects for high-precision and high-consistency welding in mass production scenarios, there are still unresolved technical problems in three sub-scenarios: environmental interference adaptation, multi-weld seam collaborative synchronization, and online welding quality control. These are all scenario-specific problems, not macro-level technical challenges, as detailed below: Photovoltaic glass welding suffers from weak anti-interference capabilities due to the lack of adaptive compensation for environmental interference: In mass production scenarios, the welding environment is prone to various interference factors such as temperature fluctuations, airflow interference, and dust pollution. These interferences directly affect the energy stability of the femtosecond laser and the welding interface state of the photovoltaic glass. Existing technologies only adopt a crude "fixed laser parameters - passive protection" approach, lacking precise compensation for environmental interference and laser parameters, and failing to establish a mechanism for predicting environmental interference. This results in unstable welding energy, poor weld formation, and defects such as porosity and cracks under environmental interference, failing to meet the stable welding requirements of mass production scenarios.

[0003] Multi-weld collaborative welding lacks synchronous control, resulting in poor weld formation consistency: Many types of photovoltaic modules (such as large-size double-glass modules) have multiple parallel or intersecting welds, requiring multi-laser head collaborative welding to improve production efficiency. However, the synchronicity of multi-weld welding directly determines the weld formation consistency. Existing technologies only adopt the modes of "single laser head sequential welding" or "multiple laser heads independent welding with fixed parameters," without multi-weld collaborative synchronous control or conflict-free planning of multi-laser head paths. This leads to large synchronization errors in multi-weld welding, easy path conflicts, and inconsistent weld formation, affecting the structural stability of photovoltaic modules.

[0004] The welding quality control is lagging behind and lacks a closed loop, resulting in a high defect rate: The welding quality of photovoltaic glass needs to be evaluated through multiple dimensions such as weld formation, mechanical properties, and light transmittance. However, the existing technology only adopts an offline control mode of "post-weld sampling inspection". There is no online real-time assessment of welding quality and parameter feedback optimization, nor has an online identification and emergency handling mechanism for welding defects been established. As a result, welding quality defects cannot be detected and corrected in a timely manner, leading to a high defect rate in mass production. Furthermore, serious defects can easily cause a large number of products to be scrapped, resulting in low production efficiency and high costs.

[0005] Existing methods for femtosecond laser welding of photovoltaic glass lack core innovations in areas such as adaptive compensation for environmental interference, collaborative synchronous control of multiple weld seams, and online closed-loop management of welding quality. Significant technological gaps exist, particularly in environmental interference-laser parameter compensation modeling, multi-weld seam synchronous control modeling, and real-time welding quality assessment-feedback modeling. There is an urgent need for a novel femtosecond laser welding method for photovoltaic glass, focusing on environmental adaptation, multi-weld seam collaboration, and closed-loop quality management. This method would achieve environmental adaptation, multi-weld seam collaboration, and closed-loop quality management in femtosecond laser welding of photovoltaic glass, filling existing technological gaps and avoiding overlap with existing photovoltaic glass welding techniques. Summary of the Invention

[0006] In response to the specific technical problems raised in the background art, the purpose of this invention is to provide an adaptive and precise control method and system for femtosecond laser welding of photovoltaic glass, so as to realize the adaptive and precise control of the entire process of femtosecond laser welding of photovoltaic glass, improve the adaptability of welding environment, the consistency of multi-weld formation, and the stability of welding quality, and further improve the technical system of high-precision welding of photovoltaic glass in mass production scenarios.

[0007] The specific technical solution of the present invention is as follows: (a) Environmental Interference Adaptive Compensation Module The core of this module is to achieve quantitative and precise compensation of the interference characteristics of the photovoltaic glass welding environment and laser parameters, as well as real-time monitoring and prediction of welding environment interference. It constructs environmental interference-laser parameter compensation modeling and environmental interference prediction modeling to solve the problems of weak anti-interference ability of the welding environment and welding quality fluctuation caused by environmental changes. It improves the adaptability and quality stability of the welding environment, and lays the foundation for stable welding in mass production scenarios.

[0008] Abandoning the traditional crude approach of "fixed parameters - passive protection," this paper constructs an integrated modeling logic encompassing environmental interference acquisition, feature extraction, compensation model construction, adaptability calculation, parameter optimization, real-time monitoring, and predictive compensation. Combining the interference characteristics of the welding environment in mass production scenarios with the welding requirements of photovoltaic glass, a welding environment interference feature extraction model, an environmental interference-laser parameter compensation correlation model, and an environmental interference prediction model are established. Furthermore, the paper designs photovoltaic glass welding environment interference feature-laser parameter compensation, and dynamic monitoring and prediction of the welding environment to achieve precise compensation and early prediction of environmental interference, ensuring stable welding quality.

[0009] First, a multi-sensor monitoring system is deployed to collect the interference characteristics of the welding environment (temperature fluctuation range, airflow speed and direction, dust concentration) and weld formation requirements (defect rate, uniformity of formation), extracting the core characteristic parameters of environmental interference. Based on laser energy transmission theory and welding experimental data, controllable parameters for laser parameter compensation (power compensation, pulse width adjustment, repetition frequency correction) are identified, and an environmental interference-laser parameter compensation correlation model is constructed. Interference compensation is designed, and the fit is calculated using the interference compensation fit formula. The compensation parameters are iteratively optimized until the fit reaches a set threshold. A dynamic monitoring model for the welding environment is constructed, and monitoring and prediction are designed. Environmental interference data is collected in real time, and the types and trends of interference are identified. A time-series prediction model is established based on historical interference data to achieve early prediction and pre-compensation of environmental interference. A compensation and prediction verification model is constructed to quantify the welding quality stability under environmental interference, dynamically optimize parameters, and ensure that the environmental interference compensation fit is ≥ the set threshold, reducing the welding quality fluctuation amplitude to within 5%.

[0010] 1: Environmental Interference Characteristics of Photovoltaic Glass Welding - Laser Parameter Compensation To address the issues of existing technologies such as "lack of quantitative compensation relationship between environmental interference and laser parameters, and weak anti-interference ability," an integrated modeling logic is constructed, which includes interference extraction, parameter screening, correlation modeling, adaptability calculation, and iterative optimization. An environmental interference-laser parameter compensation correlation model and a compensation adaptability calculation model are established, and compensation is designed to achieve quantitative and accurate compensation of environmental interference and laser parameters. This solves the problem of welding quality fluctuations caused by environmental interference and improves the adaptability of the welding environment.

[0011] Based on extensive experimental data on welding under environmental interference, this study explores the inherent coupling relationship between welding environmental interference characteristics (temperature fluctuation, airflow velocity, dust concentration) and laser parameter compensation (power compensation percentage, pulse width adjustment coefficient, repetition frequency correction value), establishing an environmental interference-laser parameter compensation correlation model. Real-time environmental interference characteristic parameters and an initial set of laser parameter compensation parameters are extracted. The compensation effect is quantified using an interference compensation fit calculation formula. Compensation parameters with fit below the threshold are iteratively optimized, adjusting parameters such as power compensation and pulse width adjustment until the fit meets the set requirements. A compensation verification model is constructed, and welding quality stability under different environmental interferences is tested through small-scale welding experiments. The correlation model and parameters are dynamically optimized to ensure that welding quality fluctuations are reduced to within 5% under various environmental interferences such as temperature fluctuations and airflow interference, improving environmental adaptability by more than 80% compared to traditional methods.

[0012] 2: Dynamic monitoring and prediction of the welding environment To address the problems of existing technologies that are "only passively protected, lack environmental interference prediction, and are slow to respond to sudden interference," an integrated modeling logic is constructed that integrates dynamic monitoring, interference identification, trend analysis, prediction modeling, and pre-compensation. A dynamic monitoring model and a time-series prediction model for welding environmental interference are established, and monitoring and prediction are designed to achieve real-time monitoring, trend prediction, and early compensation of environmental interference, thereby solving the problem of welding defects caused by sudden interference.

[0013] A dynamic monitoring model for welding environmental interference was constructed, deploying multiple types of sensors such as temperature, airflow, and dust to collect environmental interference data in real time at a sampling frequency ≥10Hz to ensure the real-time nature of the interference data. Interference identification was designed, achieving accurate identification of interference types such as temperature fluctuations, airflow interference, and dust pollution through feature extraction and classifier training. Based on historical environmental interference data, a time-series prediction model was established to predict the trend and peak value of interference changes in the next 5-10 seconds. Pre-compensation was designed, adjusting the laser parameter compensation amount in advance based on the prediction results to achieve pre-compensation for environmental interference and avoid the impact of sudden interference on welding quality. A monitoring and prediction verification model was constructed to quantify the accuracy of interference identification and prediction, dynamically optimizing parameters to ensure an interference identification accuracy ≥98%, a prediction accuracy ≥95%, and a reduction of welding defect rate caused by sudden interference by more than 90%.

[0014] (II) Multi-weld seam collaborative synchronous control module The core of this module is to achieve micron-level synchronous control of femtosecond laser multi-weld welding and conflict-free collaborative planning of multi-laser head paths. It constructs multi-weld synchronous control modeling and multi-laser head path collaborative planning modeling to solve the problems of poor synchronization, path conflict and poor forming consistency of multi-weld welding, improve the welding efficiency and forming consistency of multi-weld welding, and adapt to the multi-weld welding needs of mass production scenarios.

[0015] Abandoning the traditional crude approach of "independent welding - no synchronous planning", we construct an integrated modeling logic that integrates multi-weld feature acquisition, synchronous constraint analysis, collaborative model construction, synchronous control, path planning, conflict detection, and optimization verification. Combining the spatial distribution characteristics of multiple welds in photovoltaic glass with synchronous welding constraints, we establish a multi-weld spatial distribution-synchronous welding correlation model and a multi-laser head path conflict detection model. We design femtosecond laser multi-weld collaborative synchronous control and multi-laser head path collaborative planning to achieve precise synchronization of multiple welds and conflict-free collaborative operation of multiple laser heads.

[0016] First, a laser 3D scanner is deployed to collect the spatial distribution characteristics (weld position, length, spacing, and angle) of multiple weld seams in photovoltaic glass and the synchronous welding constraints (synchronization error, welding speed range, and laser head movement limits) to construct a spatial distribution model of multiple weld seams. Then, a multi-weld seam collaborative synchronous control model is designed, and a multi-weld welding synchronous error calculation model is established. Welding progress and forming data of each weld seam are collected in real time to calculate the synchronous error. Through coordinated fine-tuning of laser power and welding speed, micron-level synchronous control of multi-weld welding is achieved, with the synchronous error controlled within 3μm. Next, a multi-laser head path collaborative planning model is designed, employing a distributed collaborative strategy to plan the welding path and start / stop sequence of each laser head. A path conflict detection model is established to detect path conflict risks in real time and make local adjustments to conflicting paths, achieving conflict-free collaborative operation of multiple laser heads. Finally, a multi-weld seam collaborative verification model is constructed to quantify the multi-weld seam synchronization error and path conflict rate, dynamically optimizing parameters to ensure that the synchronization error is ≤3μm and the path conflict rate is reduced to 0%.

[0017] 3: Femtosecond laser multi-weld seam collaborative synchronous control To address the problems of existing technologies such as "lack of synchronous control for multiple weld seams, large synchronous errors, and poor forming consistency," an integrated modeling logic is constructed, which includes multi-weld seam feature modeling, synchronous error calculation, collaborative control, and closed-loop feedback. A spatial distribution-synchronous welding correlation model and a synchronous closed-loop control model for multiple weld seams are established. Control is designed to achieve micron-level synchronous control of multi-weld seam welding, thus solving the problem of poor forming consistency of multiple weld seams.

[0018] Based on multi-weld welding experimental data, this study explores the intrinsic correlation between the spatial distribution (spacing, angle) of multiple welds and synchronization errors, establishes a multi-weld spatial distribution-synchronous welding correlation model, and clarifies synchronization control strategies under different distribution characteristics. A synchronization error calculation method is designed, using high-speed visual inspection to collect the welding progress (welding length, penetration depth, and weld width) of each weld in real time, and calculating the synchronization error between welds. A collaborative synchronization control model is constructed, employing a "main weld reference - secondary weld follower" control strategy. Taking one key weld as the reference, precise synchronization between the secondary and main welds is achieved by fine-tuning the laser power (affecting penetration depth and welding speed) and welding speed of the secondary welds. A closed-loop feedback system is designed to feed real-time synchronization error data back to the control model, iteratively optimizing control parameters to ensure that the synchronization error of multiple welds is controlled within 3μm, improving the forming consistency by more than 90% compared to the traditional method.

[0019] 4: Multi-laser head path collaborative planning To address the problems of "lack of collaborative planning and easy path conflicts in existing technologies for multiple laser heads", an integrated modeling logic of multi-weld path modeling, conflict detection, distributed planning and time-series optimization is constructed. A multi-laser head path conflict detection model and a distributed collaborative planning model are established, and a planning is designed to achieve conflict-free collaborative planning of multi-laser head paths, thereby solving the problem of low welding efficiency caused by path conflicts.

[0020] A multi-weld seam path spatial model was constructed, generating initial welding paths for each laser head based on the spatial distribution characteristics of the weld seams. A path conflict detection system was designed, employing a spatial geometric collision detection method to detect the spatial and temporal conflict risks of each laser head path in real time, marking conflict areas and conflict time points. A distributed collaborative planning model was established, adopting a "local path adjustment-temporal optimization" strategy to locally replan conflicting paths while optimizing the start-stop timing of each laser head to avoid temporal conflicts. Collaborative planning was designed to achieve collaborative optimization of multi-laser head paths and timing, ensuring no spatial or temporal conflicts during the welding process. A path planning verification model was constructed, using simulation and actual welding experiments to detect the path conflict rate, dynamically optimizing planning parameters to ensure the path conflict rate is reduced to 0%, resulting in a multi-laser head welding efficiency improvement of over 80% compared to the traditional independent planning mode.

[0021] (III) Online Closed-Loop Control Module for Welding Quality The core of this module is to realize real-time assessment and dynamic parameter optimization of photovoltaic glass welding quality, as well as online identification and emergency handling of welding defects. It constructs a real-time welding quality assessment-feedback model and an online defect identification-emergency control model to solve the problems of lagging welding quality control and high defect rate, realize online closed-loop optimization of welding quality, and ensure the stability of welding quality in mass production scenarios.

[0022] Abandoning the traditional crude approach of "offline detection - no closed-loop feedback," this paper constructs an integrated modeling logic encompassing quality feature acquisition, real-time evaluation, parameter optimization, defect identification, emergency control, and closed-loop verification. Combining the core indicators of photovoltaic glass welding quality with the requirements of mass production, it establishes a real-time welding quality evaluation index system, a real-time welding defect identification feature library, and a quality-parameter feedback optimization model. This allows for the design of real-time evaluation of photovoltaic glass welding quality, parameter feedback optimization, online identification of welding defects, and emergency control, achieving online closed-loop management of welding quality and emergency handling of defects.

[0023] First, high-speed visual inspection and infrared thermography equipment are deployed to collect real-time quality characteristic data of the welding process (weld width, penetration depth, surface smoothness, and temperature field distribution). A real-time welding quality evaluation index system is constructed, including three primary indicators: weld formation, fusion degree, and temperature stability, with several secondary indicators. A real-time quality evaluation is designed, and a weighted scoring method is used to classify welding quality in real-time (excellent, good, medium, poor). A quality-parameter feedback optimization model is established, and parameter feedback optimization is designed to dynamically adjust laser parameters and welding processes based on the quality evaluation results, achieving closed-loop optimization of the quality-parameter parameters. Establish a real-time feature library for welding defects, labeling the visual and temperature characteristics of defects such as microcracks, porosity, and lack of fusion; design online defect identification, extracting weld features through machine vision and temperature field analysis, and matching them with the defect feature library to achieve accurate identification of defect type and severity; design emergency control, initiating emergency shutdown and parameter reset for severe defects, and adjusting parameters in real time for minor defects; construct a quality closed-loop management verification model, quantifying the welding quality pass rate and defect emergency handling success rate, dynamically optimizing parameters to ensure that the welding quality defect rate is reduced to below 1% and the defect emergency handling success rate is ≥99%.

[0024] 5: Real-time evaluation of photovoltaic glass welding quality - parameter feedback optimization To address the issues of existing technologies such as "offline detection of welding quality and lack of real-time feedback optimization," an integrated modeling logic is constructed, encompassing quality feature acquisition, evaluation modeling, scoring calculation, parameter optimization, and closed-loop feedback. This establishes a real-time welding quality evaluation index system and a quality-parameter feedback optimization model, designing evaluation and optimization mechanisms to achieve real-time evaluation of welding quality and dynamic parameter optimization, thereby resolving the problem of lagging welding quality control.

[0025] Based on the core requirements of photovoltaic glass welding quality, a real-time welding quality evaluation index system was constructed. The analytic hierarchy process (AHP) was used to set the weights of each index (weld formation weight 0.4, fusion degree weight 0.3, temperature stability weight 0.3). A quality feature acquisition system was designed, using high-speed visual inspection and infrared thermography to collect data on each index in real time, followed by standardized processing. A real-time quality evaluation model was constructed, using a weighted scoring method to calculate the comprehensive welding quality score, classifying the quality into four levels: excellent, good, medium, and poor. A quality-parameter feedback optimization model was established to explore the inherent correlation between different quality levels and laser parameters (power, pulse width, welding speed), designing parameter optimization. When the quality level is "medium" or "poor," the laser parameters are automatically adjusted, with the optimization direction being to improve fusion degree and weld formation. A closed-loop feedback model was constructed to feed the optimized quality data back to the evaluation model in real time, iteratively optimizing the evaluation parameters and laser parameters to ensure that the welding quality remains stable at the "excellent" or "good" level, improving the quality pass rate by more than 90% compared to traditional methods.

[0026] 6: Online identification and emergency control of welding defects To address the issues of existing technologies such as "inability to identify welding defects online and delayed emergency response," an integrated modeling logic is constructed, encompassing defect feature acquisition, database matching, type determination, and emergency control. This establishes a real-time welding defect identification feature database and an emergency control model, and designs identification and control mechanisms to achieve online identification and graded emergency response for welding defects, thus resolving the problem of severe defects leading to the scrapping of batches of products.

[0027] Visual features (shape, size, grayscale value) and temperature features (abnormal temperature distribution areas) of common defects such as microcracks, porosity, and lack of fusion were collected through extensive welding experiments to establish a standardized real-time welding defect identification feature library. Defect feature extraction was designed by acquiring visual images and temperature field data of weld seams in real time using high-speed visual inspection and infrared thermography equipment to extract defect feature parameters. Defect identification was designed by matching the extracted features with the feature library based on similarity; a match of ≥95% was used to identify the corresponding defect, and the severity (minor, moderate, severe) was determined based on the defect size and impact range. An emergency control model was constructed, and emergency control measures were designed. For minor defects, laser parameters were adjusted in real time (e.g., increasing power to eliminate lack of fusion); for moderate defects, welding of the corresponding weld seam was paused and locally corrected; for severe defects, the entire machine was immediately shut down to prevent defect expansion and batch product scrapping. An identification and control verification model was constructed to quantify the defect identification accuracy and emergency handling success rate, dynamically optimizing the feature library and control parameters to ensure a defect identification accuracy of ≥98% and an emergency handling success rate of ≥99%.

[0028] Beneficial effects The six core technologies of this invention are innovative technologies not reported in the prior art, each achieving a precise breakthrough in a specific technical problem. Furthermore, the three core modules work synergistically to achieve fully adaptive and precise control of the entire process of femtosecond laser welding of photovoltaic glass. Compared with existing technologies, it has the following core beneficial effects: Photovoltaic glass welding environment interference characteristics - laser parameter compensation: Abandoning the passive protection approach, a quantitative compensation model of environmental interference and laser parameters is constructed. Precise compensation is achieved through adaptability calculation, which solves the problem of weak environmental anti-interference ability, improves the adaptability of welding environment, and reduces the welding defect rate caused by environmental interference. Dynamic monitoring and prediction of welding environment: Real-time monitoring, trend prediction and pre-compensation of environmental interference, breaking through the limitations of traditional passive response, and ensuring stable welding quality in mass production scenarios; Femtosecond laser multi-weld seam collaborative synchronous control: achieves micron-level synchronous control of multi-weld seam welding, solves the problems of poor synchronization and poor forming consistency of multi-weld seams, controls the synchronization error of multi-weld seams within 3μm, improves forming consistency, and fills the technical gap in modeling of multi-weld seam synchronous control of photovoltaic glass. Multi-laser head path collaborative planning: Achieve conflict-free collaborative planning of multi-laser head paths, solve the problem of low efficiency caused by path conflicts, reduce the path conflict rate, and adapt to the high-efficiency welding requirements of mass production scenarios. Real-time assessment of photovoltaic glass welding quality - parameter feedback optimization: realizes online real-time assessment of welding quality and closed-loop optimization of parameters, solves the problem of lagging quality control, improves welding quality pass rate and reduces mass production defect rate; Online identification and emergency control of welding defects: This technology enables online identification and graded emergency handling of welding defects, solving the problem of mass scrapping caused by serious defects, significantly reducing production losses, and filling the technical gap in online emergency control modeling of welding defects in photovoltaic glass.

[0029] It achieves environmental adaptability, multi-weld seam collaboration, and quality closed-loop in femtosecond laser welding of photovoltaic glass, improving the welding yield rate compared to existing technologies. It is suitable for single-weld and multi-weld welding needs of various types of photovoltaic glass such as ultra-white, tempered, and flexible glass in mass production scenarios, increasing production efficiency by more than 80%, and has extremely strong industrial application value. Attached Figure Description

[0030] Figure 1 : Workflow diagram of the environmental interference adaptive compensation module Figure 2 : Workflow diagram of the multi-weld seam collaborative synchronous control module Detailed Implementation

[0031] The following four specific embodiments illustrate in detail the implementation steps of the present invention for photovoltaic glass welding in different scenarios, covering the full application scope of the present invention.

[0032] Example 1: Welding for mass production of flexible photovoltaic glass Implementation steps Step 1: Environmental Interference and Demand Collection: Deploy a multi-sensor monitoring system to collect interference characteristics (temperature fluctuations, airflow velocity, dust concentration) of the welding environment in the mass production of flexible photovoltaic glass, and extract core characteristic parameters; analyze the mass welding process requirements (defect rate ≤1%, forming uniformity ≥95%), and set the interference compensation adaptation threshold. .

[0033] Step 2: Interference compensation parameter modeling: Extract the initial set of controllable parameters for laser parameter compensation (power compensation, pulse width adjustment, repetition frequency correction), and construct an environmental interference-laser parameter compensation correlation model based on the environmental interference-compensation law of flexible photovoltaic glass welding.

[0034] Step 3: Fit Calculation and Parameter Optimization: Fit is calculated using the interference compensation formula. The initial fit is calculated. If the initial parameter fit is less than 93%, the compensation parameters are iteratively optimized (the power compensation is adjusted for temperature fluctuations, and the welding speed compensation is adjusted for airflow interference) until the interference compensation fit is ≥93%, and then the laser parameter compensation set is output.

[0035] Step 4: Environmental monitoring and pre-compensation: Initiate dynamic monitoring and prediction of the welding environment, monitor the interference dynamics of the mass production environment in real time, predict the trend of interference changes through the time-series prediction model, adjust the laser parameters in advance for pre-compensation, and combine real-time compensation to ensure the stability of the welding process.

[0036] Step 5: Multi-weld seam collaborative synchronous welding: For the multiple parallel weld seams of flexible photovoltaic glass, multi-weld seam collaborative synchronous control is achieved through femtosecond laser multi-weld seam synchronous control to achieve micron-level synchronous control of multiple weld seams. Through multi-laser head path collaborative planning, conflict-free paths are planned to perform batch multi-weld seam welding.

[0037] Step 6: Online Closed-Loop Quality Control: Activate the online closed-loop quality control module for welding. Through real-time quality assessment and parameter feedback optimization, the welding quality is assessed in real time and parameters are dynamically adjusted. Defects are identified online and handled in a graded manner through online defect identification and emergency control. After batch welding is completed, the welding defect rate of flexible photovoltaic glass is ≤0.8%, and the synchronous error of multiple welds is ≤2μm, which meets the requirements of batch production process.

[0038] Abandoning the traditional crude modeling approach of "fixed parameters - passive protection - offline detection," this embodiment constructs an integrated closed-loop modeling logic encompassing environmental interference quantitative compensation, real-time monitoring and prediction, multi-weld seam synchronous control, and online quality control. It takes the core requirements of flexible photovoltaic glass mass production, complex environmental interference, multi-weld seam synchronization, and low defect rate as input, breaking through the technical bottlenecks of flexible photovoltaic glass mass welding. Environmental interference compensation modeling achieves precise quantitative matching between interference and laser parameters, solving the problem of weak environmental anti-interference capability; monitoring and prediction modeling enables proactive response to interference, solving the problem of defects caused by sudden interference; multi-weld seam synchronous modeling achieves micron-level synchronization of multiple weld seams, solving the problem of poor forming consistency; and quality closed-loop modeling enables real-time quality control and emergency defect handling, solving the problem of high defect rate in mass production. This embodiment's modeling approach focuses on adaptive and precise control of flexible photovoltaic glass mass production, completely different from the crude modeling approach of existing technologies, filling the technical gap in full-process modeling of flexible photovoltaic glass mass welding.

[0039] The characteristics of environmental interference in photovoltaic glass welding – laser parameter compensation, through quantitative compensation adaptability calculation, achieves precise compensation of environmental interference and laser parameters compared to the traditional passive protection mode, improving the environmental adaptability of flexible photovoltaic glass welding and reducing the defect rate caused by environmental interference; dynamic monitoring and prediction of the welding environment, through real-time monitoring and trend prediction, ensures the continuity of mass production; femtosecond laser multi-weld seam collaborative synchronous control, through micron-level synchronization error calculation and control, multi-weld seam synchronization error ≤2μm, improving forming consistency by more than 95%, completely solving the problem of multi-weld seam synchronization difference; multi-laser head path collaborative planning, through conflict-free path planning; real-time quality assessment – ​​parameter feedback optimization, through online assessment and closed-loop optimization, improves the welding quality pass rate; online defect identification and emergency control, through online identification and graded processing.

[0040] Existing technologies for mass production of flexible photovoltaic glass lack environmental interference compensation and prediction, leading to significant fluctuations in welding quality due to environmental variations. Multiple welds are welded independently, resulting in large synchronization errors and poor consistency in forming. Offline quality inspection fails to detect defects in a timely manner, and severe defects can easily lead to the scrapping of entire batches, failing to meet the demands of mass production of flexible photovoltaic modules. The collaborative application achieves environmental adaptation, multi-weld seam collaboration, and closed-loop quality control in the mass production of flexible photovoltaic glass, completely resolving all the pain points of existing technologies. Furthermore, the modeling approach and innovations are entirely new technical points, with no overlap with existing technologies.

[0041] Example 2: Welding of multiple intersecting seams in tempered photovoltaic glass Implementation steps Step 1: Adaptive Compensation for Environmental Interference: Collect interference characteristics of the welding environment of multiple intersecting welds on tempered photovoltaic glass, and set an interference compensation adaptation threshold. The laser parameter compensation method achieves quantitative compensation of environmental interference and laser parameters by using the characteristics of photovoltaic glass welding environment interference and outputs a laser parameter compensation set; it also initiates dynamic monitoring and prediction of the welding environment, monitors and predicts environmental interference in real time, and achieves synergy between pre-compensation and real-time compensation.

[0042] Step 2: Multi-intersecting weld collaborative planning: Collect the spatial distribution characteristics (intersection angle, spacing) and synchronous welding constraints of multiple intersecting welds of tempered photovoltaic glass. Through multi-laser head path collaborative planning, plan the welding path and start-stop sequence of multiple laser heads, avoid path conflicts in the intersection area, and output a conflict-free collaborative path scheme.

[0043] Step 3: Multi-weld seam synchronous control: Start femtosecond laser multi-weld seam collaborative synchronous control, take the core intersection of the cross weld seam as the benchmark, calculate the welding progress and synchronization error of each weld seam in real time, and achieve micron-level synchronous control of multiple cross weld seams by fine-tuning the laser power and welding speed, with a synchronization error ≤3μm, and perform multi-laser head collaborative welding.

[0044] Step 4: Online monitoring of welding quality: During the welding process, the quality characteristic data of each weld is collected in real time through real-time welding quality assessment and parameter feedback optimization, the quality level is evaluated and the welding parameters are dynamically adjusted to ensure that the quality is stable at the "excellent" or "good" level.

[0045] Step 5: Online Defect Identification and Handling: Through online identification and emergency control of welding defects, weld defects are identified in real time. If a minor porosity defect is detected, the laser power is adjusted in real time for correction. If there are no serious defects, welding continues.

[0046] Step 6: Welding effect verification: After welding is completed, the synchronous error of the multiple cross welds of the tempered photovoltaic glass is ≤2.5μm, the weld formation is uniform, there are no serious defects such as cracks or lack of fusion, and the bonding strength of the cross area is ≥28MPa, which meets the process requirements of the multiple cross welds of the tempered photovoltaic glass.

[0047] Abandoning the traditional crude modeling approach of "independent welding - no synchronization - offline detection," this embodiment constructs an integrated closed-loop modeling logic encompassing environmental interference compensation, multi-crossing path collaboration, multi-weld synchronous control, and online quality control. It takes the core requirements of tempered photovoltaic glass—multiple cross-welds, potential path conflicts, high synchronization requirements, and high bond strength requirements—as input, breaking through the technical bottlenecks in welding multiple cross-welds of tempered photovoltaic glass. Environmental interference compensation modeling achieves stable welding quality under complex environments, solving the problem of defects in cross-area areas caused by environmental interference. Path collaboration modeling enables conflict-free path planning for multiple cross-welds, solving the problem of low efficiency caused by path conflicts. Synchronous control modeling achieves micron-level synchronization of multiple welds, solving the problem of poor forming in cross-area areas. Quality closed-loop modeling enables real-time quality control and defect correction, solving the problem of insufficient bond strength in cross-area areas. This embodiment focuses on the collaborative and precise control of multiple cross-welds in tempered photovoltaic glass, completely different from the decentralized modeling approach of existing technologies, filling the technical gap in integrated modeling of multiple cross-weld welding in tempered photovoltaic glass.

[0048] The collaborative application of laser parameter compensation and monitoring prediction in the welding environment of photovoltaic glass enables online defect identification and emergency control, which reduces the defect rate by correcting minor defects in real time. The overall collaborative operation ensures that the forming consistency and bonding strength of the multi-intersecting welds of tempered photovoltaic glass meet the process requirements, which is an improvement over existing technologies.

[0049] Existing technologies for welding multiple intersecting welds in tempered photovoltaic glass lack environmental interference compensation, making the intersecting areas prone to defects due to environmental fluctuations. Multiple laser heads independently plan paths, resulting in high path conflict rates and low welding efficiency in the intersecting areas. The lack of synchronous control across multiple welds leads to large synchronization errors in the intersecting areas, poor weld formation, and insufficient bonding strength. Offline quality inspection prevents timely defect correction, resulting in a weld yield of less than 60% for multiple intersecting welds, failing to meet the structural stability requirements of tempered photovoltaic modules. Our collaborative application achieves conflict-free, high-precision, synchronous, and high-quality welding of multiple intersecting welds in tempered photovoltaic glass, completely resolving the pain points of existing technologies. Weld yield is improved, and the bonding strength and consistency of the weld formation meet the requirements of high-end tempered photovoltaic modules. Furthermore, the modeling approach and innovations are entirely new technical points, adaptable to the industrial welding needs of multiple intersecting welds.

[0050] Example 3: Batch single-seam edge sealing welding of ultra-white photovoltaic glass Implementation steps Step 1: Precise Environmental Interference Compensation: Collect interference characteristics (temperature fluctuation, dust concentration) of the welding environment for single weld seams of ultra-white photovoltaic glass in batches, and set the interference compensation adaptation threshold. The interference and laser parameters are quantitatively compensated by the interference characteristics of the photovoltaic glass welding environment and laser parameter compensation; dynamic monitoring and prediction of the welding environment are initiated to monitor environmental changes in real time and pre-compensate in advance to ensure the adaptability of the welding environment.

[0051] Step 2: Welding Path Planning and Welding: For the straight and arc transition characteristics of batch single weld seam sealing, the optimal welding path of a single laser head is planned through multi-laser head path collaborative planning to avoid idle path loss; femtosecond laser welding is started, and weld quality characteristic data is collected in real time through welding quality real-time evaluation and parameter feedback optimization to evaluate the quality level and dynamically adjust the laser parameters.

[0052] Step 3: Online Defect Monitoring and Handling: During the welding process, the weld is monitored in real time through online identification and emergency control of welding defects. If a medium-sized non-fusion defect is found, the welding of the product is immediately suspended. After adjusting the laser power, local repair welding is carried out. After the repair, batch welding continues.

[0053] Step 4: Batch welding quality control: Continuously conduct real-time quality assessment and defect monitoring for each product produced in batches through the online closed-loop quality control module to ensure that the quality of each product meets the standards.

[0054] Step 5: Batch production effect verification: Complete the batch single-seam edge sealing welding of 1000 pieces of ultra-white photovoltaic glass. The overall welding defect rate is ≤0.5%, the weld width uniformity is ≥98%, the light transmittance loss is ≤0.8%, and the edge sealing performance reaches IP68 level, which meets the process requirements for batch edge sealing welding of ultra-white photovoltaic glass.

[0055] Abandoning the traditional extensive modeling approach of "fixed parameters - manual inspection - offline testing," this embodiment constructs an integrated closed-loop modeling logic encompassing environmental interference quantitative compensation, path optimization, online quality control, and defect emergency handling. It takes the core requirements of batch edge sealing of ultra-white photovoltaic glass—high sealing performance, low light transmission loss, and low defect rate—as input, breaking through the technical bottlenecks in batch edge sealing welding of ultra-white photovoltaic glass. Environmental interference compensation modeling achieves stable quality under batch production conditions, solving the problem of edge sealing defects caused by environmental fluctuations. Path optimization modeling enables efficient planning of welding paths, solving the problem of low batch production efficiency. Quality closed-loop modeling enables real-time quality control of each product, solving the problem of poor quality consistency in batch production. Defect emergency modeling enables timely defect handling, solving the problem of high defect rates in batch production. This embodiment's modeling approach focuses on the efficient and precise control of batch edge sealing of ultra-white photovoltaic glass, completely different from the existing manual control modeling approach, filling the technical gap in integrated modeling of batch edge sealing welding of ultra-white photovoltaic glass.

[0056] Photovoltaic glass welding environment interference characteristics - Laser parameter compensation: Through quantitative compensation, the environmental adaptability of ultra-white photovoltaic glass edge sealing welding is improved, and the defect rate caused by environmental interference is reduced; Multi-laser head path collaborative planning: Through optimal path planning, the efficiency of batch edge sealing welding is improved, and product scrapping due to defects is avoided; Overall collaborative operation improves the production efficiency of batch edge sealing welding of ultra-white photovoltaic glass.

[0057] Existing technologies for mass edge sealing of ultra-clear photovoltaic glass lack environmental interference compensation, leading to a defect rate of ≥12% due to environmental fluctuations during mass production. Manual path planning results in significant idle time loss and low production efficiency. Manual inspection and offline testing result in lagging quality control, poor consistency in mass production quality, unstable edge sealing, and a batch edge sealing welding yield of less than 75%, failing to meet the mass production requirements of ultra-clear photovoltaic modules. Our collaborative application achieves high efficiency, low defects, high sealing, and low light transmission loss in the mass edge sealing welding of ultra-clear photovoltaic glass, completely resolving the pain points of existing technologies. This improves the batch production yield and efficiency, and the modeling approach and innovations are entirely new technologies, adaptable to the large-scale mass production of ultra-clear photovoltaic glass.

[0058] Example 4: Integrated Application of Multi-Type Photovoltaic Glass Mixed Batch Welding Production Line Implementation steps Step 1: Full-process and module deployment: Integrate the six core components of this invention into the femtosecond laser photovoltaic glass hybrid batch welding production line, deploy three core modules: environmental interference adaptive compensation, multi-weld seam collaborative synchronous control, and online closed-loop control of welding quality, and build a full-process data interaction module to realize real-time data transmission, storage, and parameter iterative optimization of each module.

[0059] Step 2: Multi-type glass environment adaptation: For the three types of photovoltaic glass on the production line, namely flexible, tempered and ultra-clear, there are three welding scenarios, namely batch multi-weld seam, multi-intersecting weld seam and batch edge sealing. By using the photovoltaic glass welding environment interference characteristics - laser parameter compensation and dynamic monitoring and prediction of the welding environment, different interference compensation thresholds are set to achieve environmental adaptive compensation for welding of each type of glass.

[0060] Step 3: Multi-scenario collaborative planning and welding: For the three welding scenarios, conflict-free welding paths are planned separately through multi-laser head path collaborative planning; micron-level synchronization of multi-weld seam scenarios and precise synchronization of multi-intersecting weld seam scenarios are achieved through femtosecond laser multi-weld seam collaborative synchronous control; the production line is started to carry out mixed batch welding of multiple types of glass.

[0061] Step 4: Closed-loop quality control throughout the entire process: During the welding process, the welding quality of the three types of glass is evaluated in real time and the parameters are dynamically optimized through real-time welding quality assessment and parameter feedback optimization; through online identification and emergency control of welding defects, various defects are identified and graded online to ensure stable welding quality.

[0062] Step 5: Parameter Iteration and Optimization: The full-process data interaction module collects full-process data for welding three types of glass and iteratively optimizes six core parameters through machine learning to improve adaptability and accuracy for multiple types of glass and multiple welding scenarios.

[0063] Step 6: Verification of large-scale application effect: Apply the integrated application solution to the large-scale production of a mixed batch welding production line for multiple types of photovoltaic glass, continuously producing 3,000 products (1,000 of each type), with an overall welding yield of ≥99%, a synchronous error of multiple welds ≤3μm, a welding defect rate of ≤0.8%, and a production efficiency that is more than 85% higher than that of traditional production lines. The welding performance of each type of glass meets the high standard requirements for photovoltaic module production.

[0064] Abandoning the traditional extensive modeling approach of "single-type production line - independent control - no iteration," this embodiment constructs an integrated closed-loop production modeling logic encompassing multi-type environmental differentiation compensation, multi-scenario collaborative planning, multi-type quality unified closed loop, and data-driven iteration. It takes the core requirements of multi-type, multi-scenario, high consistency, and high efficiency of photovoltaic glass mixed batch welding production lines as input, breaking through the technical bottlenecks of traditional production lines' single adaptability and low efficiency. Multi-type environmental differentiation compensation modeling achieves environmental adaptability for different glasses and scenarios, solving the problem of inconsistent production line environmental adaptation. Multi-scenario collaborative planning modeling achieves efficient collaboration across different welding scenarios, solving the problem of complex production line path planning. Multi-type quality unified closed-loop modeling achieves unified control of welding quality for different glasses, solving the problem of inconsistent production line quality standards. Data-driven iterative modeling enables continuous parameter optimization, solving the problem of insufficient adaptability to multiple types and scenarios. This embodiment's modeling approach focuses on the adaptive and precise control of multi-type photovoltaic glass mixed batch welding production lines, completely different from the existing single-type production line modeling approach, filling the technical gap in integrated modeling of multi-type photovoltaic glass mixed batch welding production lines.

[0065] The differentiated application, collaborative operation, and iterative optimization on the production line, compared to the independent mode of traditional single-type production lines, have achieved full-process adaptive and precise control for multiple types of photovoltaic glass and multiple welding scenarios. The differentiated application of the environmental interference adaptive compensation module has achieved environmental adaptive compensation for welding of three types of glass, improving the environmental adaptability of each type of glass welding by more than 80% and reducing the defect rate caused by environmental interference by more than 90%. The multi-scenario application of the multi-weld seam collaborative synchronization control module has achieved precise synchronization and conflict-free path planning for multiple weld seams and multiple intersecting weld seams, with a multi-weld seam synchronization error of ≤3μm, a path conflict rate reduced to 0%, and an overall welding efficiency improvement of more than 85% for the production line. The unified application of the welding quality online closed-loop management module has achieved real-time evaluation, parameter optimization, and emergency handling of welding quality for three types of glass, with a defect rate of ≤0.8% and a yield rate of ≥99% for each type of glass welding. The iterative optimization of the full-process data interaction module has continuously improved the adaptability of multiple scenarios and improved the quality consistency of large-scale production. Overall collaborative operation has improved the production efficiency of the multi-type photovoltaic glass mixed batch welding production line by a percentage, realizing the large-scale, intelligent, and precise production of photovoltaic glass welding.

[0066] Existing photovoltaic glass welding production lines are only adaptable to welding single types of glass and single scenarios. They lack environmental differentiation compensation, multi-scenario collaborative planning, and unified quality closed-loop control. When welding multiple types of glass in batches, problems such as poor environmental adaptability, path conflicts, and inconsistent quality easily arise. Each module operates independently, data is not shared, and there is no iterative optimization. The production line has poor adaptability and low production efficiency, with a mixed batch welding yield of less than 70%, which cannot meet the needs of multi-type and large-scale photovoltaic module production. Differentiated application + collaborative operation + iterative optimization realizes full-process adaptive and precise control of the mixed batch welding production line for multiple types of photovoltaic glass. It completely solves the pain points of existing technologies. The overall welding yield of the production line is ≥99%, and the production efficiency is improved by more than 85%. Moreover, the modeling ideas and innovations are all new technical points, providing a brand-new solution for the intelligent and large-scale upgrading of photovoltaic glass welding production lines, with extremely strong industrial application value.

[0067] The foregoing has shown and described the basic principles, main features, and core advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely the innovative and modeling principles of the present invention. Various changes and modifications can be made to the present invention without departing from its spirit and core scope, and all such changes and modifications fall within the scope of protection claimed by the present invention. The scope of protection claimed by the appended claims and their equivalents is defined.

Claims

1. An adaptive precision control method for femtosecond laser welding of photovoltaic glass, characterized in that, Includes the following steps: S1: Adaptive compensation processing for environmental interference. It collects the characteristics of environmental interference and weld formation requirements for photovoltaic glass welding. Through laser parameter compensation based on the characteristics of photovoltaic glass welding environmental interference, dynamic monitoring and prediction of the welding environment, it realizes the quantitative matching of environmental interference and laser parameters, real-time monitoring and prediction compensation of environmental interference, and outputs an adapted laser parameter compensation set and environmental response strategy. S2: Multi-weld seam collaborative synchronous control processing, collects the distribution characteristics of multiple weld seams of photovoltaic glass and synchronous welding constraints, realizes micron-level synchronous control of multi-weld seam welding and conflict-free collaborative planning of multiple laser head paths through femtosecond laser multi-weld seam collaborative synchronous control and multi-laser head path collaborative planning, and outputs multi-weld seam synchronous control parameters and collaborative path planning scheme. S3: Online closed-loop control and processing of welding quality, real-time collection of quality characteristics and weld formation data of photovoltaic glass welding process, real-time evaluation of photovoltaic glass welding quality - parameter feedback optimization, online identification and emergency control of welding defects, to achieve real-time evaluation of welding quality and dynamic optimization of parameters, online identification and emergency handling of welding defects, to ensure stable and compliant welding quality; Among them, the photovoltaic glass welding environment interference characteristics - laser parameter compensation in step S1 includes the interference compensation fit calculation formula: The constraints are , Interference compensation fit (%) To compensate for the number of feature parameters, These are the core characteristic parameters of welding environment interference. The core parameters for laser parameter compensation, To compensate for the compatibility threshold, it is set according to the type of photovoltaic glass and welding requirements: flexible photovoltaic glass welding μ0≥93%, tempered photovoltaic glass welding μ0≥90%, and ultra-white photovoltaic glass welding μ0≥88%.

2. The method according to claim 1, characterized in that, The photovoltaic glass welding environment interference characteristics-laser parameter compensation in step S1 includes the following sub-steps: extracting welding environment interference characteristics and controllable parameters for laser parameter compensation, constructing an environment interference-laser parameter compensation correlation model, calculating the fit degree through the interference compensation fit degree calculation formula, iteratively optimizing the compensation parameters until the fit degree reaches the set threshold, and realizing quantitative and accurate compensation of environmental interference and laser parameters.

3. The method according to claim 1, characterized in that, The dynamic monitoring and prediction of the welding environment in step S1 includes the following sub-steps: constructing a dynamic monitoring model for welding environment interference, collecting environmental interference data in real time and identifying the type and trend of interference, establishing a prediction model based on historical interference data, realizing early prediction and pre-compensation of environmental interference, and reducing the impact of sudden interference on welding quality.

4. The method according to claim 1, characterized in that, The femtosecond laser multi-weld seam collaborative synchronous control in step S2 includes the following sub-steps: constructing a spatial distribution-synchronous welding correlation model of photovoltaic glass multi-weld seams, calculating the synchronization error of multi-weld seam welding in real time, and achieving micron-level synchronous control of multi-weld seam welding through coordinated fine-tuning of laser power and welding speed, with the synchronization error controlled within 3μm.

5. The method according to claim 1, characterized in that, The multi-laser head path collaborative planning in step S2 includes the following sub-steps: establishing a multi-laser head path conflict detection model, adopting a distributed collaborative planning strategy, planning the welding path and start-stop sequence of each laser head, realizing conflict-free collaborative operation of multiple laser heads, and reducing the path conflict rate to 0%.

6. The method according to claim 1, characterized in that, The real-time evaluation and parameter feedback optimization of photovoltaic glass welding quality in step S3 includes the following sub-steps: constructing a real-time evaluation index system for welding quality, collecting quality characteristic data of the welding process in real time, evaluating the welding quality level through a weighted scoring method, and dynamically optimizing laser parameters and welding process based on the evaluation results to achieve closed-loop feedback of quality and parameters.

7. The method according to claim 1, characterized in that, The online identification and emergency control of welding defects in step S3 includes the following sub-steps: establishing a real-time identification feature library for welding defects, acquiring weld images in real time through machine vision and extracting features, matching them with the feature library to identify the type and severity of defects, initiating emergency shutdown and parameter reset for severe defects, and adjusting and correcting parameters in real time for minor defects.

8. The method according to claim 1, characterized in that, The interference compensation adaptability threshold It can be flexibly adjusted according to the photovoltaic glass welding process scenario, with μ0≥90% in batch production scenarios and μ0≥95% in high-precision laboratory scenarios, and the compensation can adapt to various environmental interference types such as temperature fluctuations, airflow interference, and dust pollution.

9. The method according to any one of claims 1-8, characterized in that, The method can be applied to single-seam and multi-seam welding of various types of photovoltaic glass, such as ultra-white, tempered, flexible, and borosilicate, in mass production scenarios, realizing environmental adaptability, multi-seam collaboration, and quality closed-loop of femtosecond laser welding of photovoltaic glass.

10. An adaptive precision control system for femtosecond laser welding of photovoltaic glass, characterized in that, The system includes an environmental interference adaptive compensation module, a multi-weld seam collaborative synchronous control module, an online closed-loop control module for welding quality, and a full-process data interaction module. The environmental interference adaptive compensation module executes the methods described in claims 1-3 and 8. The multi-weld seam collaborative synchronous control module executes the methods described in claims 1, 4-5, and 8-9. The online closed-loop control module for welding quality executes the methods described in claims 1 and 6-9. The full-process data interaction module enables real-time data transmission, storage, and parameter iterative optimization of each module, ensuring accurate and coordinated operation of the entire system.

Images